{
 "license": "Facts are sourced; cite planlimits.dev",
 "checked": "2026-10-15",
 "plans": {
  "checked": "2026-10-15",
  "note": "Every value comes from the cited source on the checked date. 'unknown' = not stated on the source we checked.",
  "vendors": [
   {
    "id": "claude-code",
    "name": "Claude Code (Anthropic)",
    "sources": [
     "https://claude.com/pricing",
     "https://support.claude.com/en/articles/11145838-use-claude-code-with-your-pro-or-max-plan"
    ],
    "usage_model": "Rolling 5-hour session window plus weekly limits on paid plans; shared across Claude apps and Claude Code. No fixed message count.",
    "plans": [
     {
      "name": "Free",
      "price_usd_month": 0,
      "window": "5-hour",
      "weekly_cap": "not stated",
      "models": "Sonnet 5.5, Haiku (chat only)",
      "overage": "No usage credits",
      "coding_agent": "No Claude Code",
      "source": "https://claude.com/pricing"
     },
     {
      "name": "Pro",
      "price_usd_month": 20,
      "price_note": "$17/mo billed annually ($200 up front)",
      "window": "5-hour (at least 5x Free per session)",
      "weekly_cap": "Yes (amount not published)",
      "models": "Opus, Sonnet, Haiku; Fable via usage credits",
      "overage": "Usage credits at standard API rates",
      "coding_agent": "Claude Code included",
      "source": "https://claude.com/pricing"
     },
     {
      "name": "Max 5x",
      "price_usd_month": 100,
      "price_note": "Pricing page says 'From $100'",
      "window": "5-hour (5x Pro per session)",
      "weekly_cap": "Yes (amount not published); separate weekly Fable limit (50% of weekly limits)",
      "models": "Opus, Sonnet, Haiku, Fable",
      "overage": "Usage credits at standard API rates; $100/mo API credits (not usable in Claude Code)",
      "coding_agent": "Claude Code included",
      "source": "https://claude.com/pricing"
     },
     {
      "name": "Max 20x",
      "price_usd_month": 200,
      "price_note": "$200/mo (Claude Help Center: What is the Max plan?)",
      "window": "5-hour (20x Pro per session)",
      "weekly_cap": "Yes (amount not published); separate weekly Fable limit",
      "models": "Opus, Sonnet, Haiku, Fable",
      "overage": "Usage credits; $200/mo API credits (not usable in Claude Code)",
      "coding_agent": "Claude Code included",
      "source": "https://support.claude.com/en/articles/11049741-what-is-the-max-plan"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Baseline captured by PlanLimits.",
      "source": "https://claude.com/pricing"
     },
     {
      "date": "2026-10-15",
      "change": "v3 data pass: filled gaps (models, prices, estimates).",
      "source": "https://claude.com/pricing"
     }
    ]
   },
   {
    "id": "codex",
    "name": "OpenAI Codex (ChatGPT plans)",
    "sources": [
     "https://chatgpt.com/codex/pricing/",
     "https://developers.openai.com/codex/pricing/",
     "https://www.seelig.ai/en/benchmarks/subscription-limits/chatgpt"
    ],
    "usage_model": "Local messages and cloud chats share a 5-hour window; additional weekly limits may apply. Message counts vary by model and task.",
    "plans": [
     {
      "name": "Free / Go ($8)",
      "price_usd_month": 0,
      "window": "5-hour",
      "weekly_cap": "may apply",
      "models": "GPT-6 Luna at Standard speed in the desktop app",
      "overage": "unknown",
      "coding_agent": "Codex included",
      "source": "https://chatgpt.com/codex/pricing/",
      "price_note": "Go is $8/mo"
     },
     {
      "name": "Plus",
      "price_usd_month": 20,
      "window": "5-hour; 5–45 GPT-6 Astra / 15–160 GPT-6.1 Sol / 350–3,000 GPT-6 Luna local msgs",
      "weekly_cap": "may apply",
      "models": "GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol, GPT-6 Luna",
      "overage": "Credits (usage varies by model)",
      "coding_agent": "Codex included",
      "source": "https://developers.openai.com/codex/pricing/"
     },
     {
      "name": "Pro 100",
      "price_usd_month": 100,
      "window": "No 5-hour limit; weekly allowance 5x Plus (OpenAI on X)",
      "weekly_cap": "may apply",
      "models": "GPT-6 Astra included in allowance",
      "overage": "Credits",
      "coding_agent": "Codex included",
      "source": "https://chatgpt.com/codex/pricing/"
     },
     {
      "name": "Pro 200",
      "price_usd_month": 200,
      "window": "No 5-hour limit; 10x Plus for new subs, 20x for existing until Oct 29",
      "weekly_cap": "may apply",
      "models": "GPT-6 Astra included in allowance",
      "overage": "Credits",
      "coding_agent": "Codex included",
      "source": "https://developers.openai.com/codex/pricing/"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Baseline captured. Plus/Free/Go prices not stated on the Codex pricing pages.",
      "source": "https://chatgpt.com/codex/pricing/"
     },
     {
      "date": "2026-10-15",
      "change": "Codex page now lists Plus $20 and Go $8. Plus estimates changed to 5–45 GPT-6 Astra / 15–160 GPT-6.1 Sol / 350–3,000 GPT-6 Luna per 5h. Pro: no 5-hour limit.",
      "source": "https://developers.openai.com/codex/pricing/"
     },
     {
      "date": "2026-10-15",
      "change": "v3 data pass: filled gaps (models, prices, estimates).",
      "source": "https://chatgpt.com/codex/pricing/"
     }
    ]
   },
   {
    "id": "cursor",
    "name": "Cursor",
    "sources": [
     "https://cursor.com/pricing",
     "https://cursor.com/docs/models-and-pricing"
    ],
    "usage_model": "One usage pool for Cursor models, refreshes weekly. Third-party models (Claude, GPT, Gemini) billed at API list price from prepaid credit.",
    "plans": [
     {
      "name": "Free",
      "price_usd_month": 0,
      "window": "weekly pool",
      "weekly_cap": "Limited",
      "models": "Cursor's first-party Grok models only (limited); no Auto, no third-party",
      "overage": "Third-party models not available",
      "coding_agent": "Agent",
      "source": "https://cursor.com/docs/models-and-pricing"
     },
     {
      "name": "Plus",
      "price_usd_month": 8,
      "window": "weekly pool",
      "weekly_cap": "Base (1x)",
      "models": "Grok 4.7, Grok 4.6, Grok 4.5, Composer 2.5 in the pool; no third-party models",
      "overage": "Third-party models not available",
      "coding_agent": "Agent, CLI",
      "source": "https://cursor.com/docs/models-and-pricing"
     },
     {
      "name": "Premium",
      "price_usd_month": 30,
      "window": "weekly pool",
      "weekly_cap": "4x Plus",
      "models": "Grok 4.7, Grok 4.6, Grok 4.5, Composer 2.5 in the pool; Claude/GPT/Gemini on demand at API list price",
      "overage": "Top-offs / prepaid credit",
      "coding_agent": "Agent, CLI, Cloud Agents",
      "source": "https://cursor.com/docs/models-and-pricing"
     },
     {
      "name": "Super",
      "price_usd_month": 100,
      "window": "weekly pool",
      "weekly_cap": "15x Plus",
      "models": "Grok 4.7, Grok 4.6, Grok 4.5, Composer 2.5 in the pool; Claude/GPT/Gemini on demand at API list price",
      "overage": "Prepaid credit",
      "coding_agent": "Agent, CLI, Cloud Agents",
      "source": "https://cursor.com/docs/models-and-pricing"
     },
     {
      "name": "Ultra",
      "price_usd_month": 200,
      "window": "weekly pool",
      "weekly_cap": "50x Plus",
      "models": "Grok 4.7, Grok 4.6, Grok 4.5, Composer 2.5 in the pool; Claude/GPT/Gemini on demand at API list price",
      "overage": "Prepaid credit",
      "coding_agent": "Agent, CLI, Cloud Agents",
      "source": "https://cursor.com/docs/models-and-pricing"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Baseline captured. Legacy Hobby/Pro/Pro+/Ultra plans continue for existing customers.",
      "source": "https://cursor.com/docs/models-and-pricing"
     },
     {
      "date": "2026-10-15",
      "change": "Pool models named: Grok 4.7, 4.6, 4.5, Composer 2.5.",
      "source": "https://cursor.com/docs/models-and-pricing"
     },
     {
      "date": "2026-10-15",
      "change": "v3 data pass: filled gaps (models, prices, estimates).",
      "source": "https://cursor.com/pricing"
     }
    ]
   },
   {
    "id": "copilot",
    "name": "GitHub Copilot",
    "sources": [
     "https://github.com/features/copilot/plans"
    ],
    "usage_model": "Monthly GitHub AI Credits (1 credit = $0.01). Chat, agent mode, CLI, code review consume credits; completions unlimited on paid plans. Resets monthly.",
    "plans": [
     {
      "name": "Free",
      "price_usd_month": 0,
      "window": "monthly",
      "weekly_cap": "none stated",
      "models": "Large multi-vendor list",
      "overage": "n/a",
      "coding_agent": "Agent mode, CLI (limited)",
      "allowance": "2000 completions + 50 chat requests",
      "source": "https://github.com/features/copilot/plans"
     },
     {
      "name": "Pro",
      "price_usd_month": 10,
      "window": "monthly",
      "weekly_cap": "none stated",
      "models": "Multi-vendor (Claude, GPT, Gemini, Grok, Kimi)",
      "overage": "Paid credits at $0.01 each with budget",
      "coding_agent": "Agent, cloud agent, CLI",
      "allowance": "$15/mo total credits ($10 base + $5 flex)",
      "source": "https://github.com/features/copilot/plans"
     },
     {
      "name": "Pro+",
      "price_usd_month": 39,
      "window": "monthly",
      "weekly_cap": "none stated",
      "models": "Multi-vendor",
      "overage": "Paid credits at $0.01 each",
      "coding_agent": "Agent, cloud agent, CLI",
      "allowance": "$70/mo total credits ($39 base + $31 flex)",
      "source": "https://github.com/features/copilot/plans"
     },
     {
      "name": "Max",
      "price_usd_month": 100,
      "window": "monthly",
      "weekly_cap": "none stated",
      "models": "Multi-vendor",
      "overage": "Paid credits at $0.01 each",
      "coding_agent": "Agent, cloud agent, CLI",
      "allowance": "$200/mo total credits shown ($100 base; flex amount not shown)",
      "source": "https://github.com/features/copilot/plans"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Baseline captured. Flex allotments 'may change over time' per GitHub.",
      "source": "https://github.com/features/copilot/plans"
     },
     {
      "date": "2026-10-15",
      "change": "v3 data pass: filled gaps (models, prices, estimates).",
      "source": "https://github.com/features/copilot/plans"
     }
    ]
   },
   {
    "id": "windsurf",
    "name": "Windsurf / Devin Desktop",
    "sources": [
     "https://devin.ai/blog/windsurf-pricing-plans",
     "https://docs.devin.ai/desktop/accounts/quota",
     "https://docs.devin.ai/desktop/models"
    ],
    "usage_model": "Daily + weekly token-based quota (since 2026-03-19). Extra usage billed at API list price.",
    "plans": [
     {
      "name": "Free",
      "price_usd_month": 0,
      "window": "daily + weekly",
      "weekly_cap": "Light quota",
      "models": "Limited; free in-house SWE models don't count against quota",
      "overage": "Wait for reset",
      "coding_agent": "Cascade agent",
      "source": "https://devin.ai/blog/windsurf-pricing-plans"
     },
     {
      "name": "Pro",
      "price_usd_month": 20,
      "window": "daily + weekly",
      "weekly_cap": "Increased quota",
      "models": "Full model availability incl. Claude Opus 5.5 via Fusion, GPT-6, Gemini 4 Argon, SWE-2",
      "overage": "Extra usage at API price",
      "coding_agent": "Cascade agent",
      "source": "https://devin.ai/blog/windsurf-pricing-plans"
     },
     {
      "name": "Max",
      "price_usd_month": 200,
      "window": "daily + weekly",
      "weekly_cap": "Significantly higher quota",
      "models": "Full model availability incl. Claude Opus 5.5 via Fusion, GPT-6, Gemini 4 Argon, SWE-2",
      "overage": "Extra usage at API price",
      "coding_agent": "Cascade agent",
      "source": "https://devin.ai/blog/windsurf-pricing-plans"
     },
     {
      "name": "Teams",
      "price_usd_month": 40,
      "price_note": "per seat",
      "window": "daily + weekly",
      "weekly_cap": "Pro-level",
      "models": "Full model availability incl. Claude Opus 5.5 via Fusion, GPT-6, Gemini 4 Argon, SWE-2",
      "overage": "Extra usage at API price",
      "coding_agent": "Cascade agent",
      "source": "https://devin.ai/blog/windsurf-pricing-plans"
     }
    ],
    "changelog": [
     {
      "date": "2026-03-19",
      "change": "Credits replaced by daily/weekly quotas; Pro $15 -> $20 (legacy $15 grandfathered); Teams $30 -> $40/seat; new Max $200.",
      "source": "https://devin.ai/blog/windsurf-pricing-plans"
     },
     {
      "date": "2026-10-15",
      "change": "Baseline captured. windsurf.com/pricing returned 404 to our fetcher.",
      "source": "https://docs.devin.ai/desktop/accounts/quota"
     },
     {
      "date": "2026-10-15",
      "change": "v3 data pass: filled gaps (models, prices, estimates).",
      "source": "https://devin.ai/blog/windsurf-pricing-plans"
     }
    ]
   },
   {
    "id": "gemini",
    "name": "Gemini Code Assist / Gemini CLI (Google)",
    "sources": [
     "https://cloud.google.com/products/gemini/pricing",
     "https://developers.google.com/gemini-code-assist/resources/quotas",
     "https://developers.google.com/gemini-code-assist/resources/faqs",
     "https://docs.cloud.google.com/gemini/docs/codeassist/sunset",
     "https://developers.google.com/gemini-code-assist/docs/gemini-3"
    ],
    "usage_model": "Daily request quota per user (agent mode + CLI combined). Being sunset in favor of Antigravity.",
    "plans": [
     {
      "name": "Individuals / AI Pro / AI Ultra",
      "price_usd_month": "n/a",
      "window": "n/a",
      "weekly_cap": "n/a",
      "models": "n/a",
      "overage": "n/a",
      "coding_agent": "Discontinued 2026-06-18 for IDE + Gemini CLI",
      "source": "https://developers.google.com/gemini-code-assist/resources/faqs"
     },
     {
      "name": "Standard",
      "price_usd_month": 22.8,
      "price_note": "$19 with 12-month commitment; no new purchases since 2026-10-09",
      "window": "daily",
      "weekly_cap": "1500 agent/CLI requests per user/day",
      "models": "Gemini 3.5 Flash (default, GA); Gemini 3.1 Pro on Preview channel",
      "overage": "Quota increase requests",
      "coding_agent": "Agent mode, Gemini CLI",
      "source": "https://cloud.google.com/products/gemini/pricing?hl=pt-br"
     },
     {
      "name": "Enterprise",
      "price_usd_month": 54,
      "price_note": "$45 with 12-month commitment; no new purchases since 2026-10-09",
      "window": "daily",
      "weekly_cap": "2000 agent/CLI requests per user/day",
      "models": "Gemini 3.5 Flash (default, GA); Gemini 3.1 Pro on Preview channel",
      "overage": "Quota increase requests",
      "coding_agent": "Agent mode, Gemini CLI",
      "source": "https://cloud.google.com/products/gemini/pricing?hl=pt-br"
     }
    ],
    "changelog": [
     {
      "date": "2026-06-18",
      "change": "Consumer tiers stop serving Code Assist IDE + Gemini CLI; migrate to Antigravity.",
      "source": "https://developers.google.com/gemini-code-assist/resources/faqs"
     },
     {
      "date": "2026-10-09",
      "change": "New Standard/Enterprise subscriptions no longer sold.",
      "source": "https://docs.cloud.google.com/gemini/docs/codeassist/sunset"
     },
     {
      "date": "2026-10-15",
      "change": "v3 data pass: filled gaps (models, prices, estimates).",
      "source": "https://cloud.google.com/products/gemini/pricing"
     }
    ]
   },
   {
    "id": "zai",
    "name": "Z.ai GLM Coding Plan",
    "sources": [
     "https://docs.z.ai/devpack/overview",
     "https://www.aicompare.ninja/en/tool/zhipu-glm",
     "https://docs.z.ai/devpack/transition"
    ],
    "usage_model": "Credits per rolling 5h + weekly; off-peak usage costs 50% of credits.",
    "plans": [
     {
      "window": "2,000 credits / 5h",
      "weekly_cap": "10,000 credits / week",
      "overage": "—",
      "coding_agent": "Claude Code, OpenCode, Cline, ZCode…",
      "name": "Lite",
      "price_usd_month": 18,
      "models": "GLM-5.3, GLM-5.3-Flash",
      "source": "https://docs.z.ai/devpack/overview"
     },
     {
      "window": "12,000 credits / 5h",
      "weekly_cap": "60,000 / week",
      "overage": "—",
      "coding_agent": "same",
      "name": "Pro",
      "price_usd_month": 80,
      "price_note": "$80 per AI Compare Ninja (Oct 2026); earlier standard $72",
      "models": "GLM-5.3, GLM-5.3-Flash",
      "source": "https://www.aicompare.ninja/en/tool/zhipu-glm"
     },
     {
      "window": "28,000 credits / 5h",
      "weekly_cap": "140,000 / week",
      "overage": "—",
      "coding_agent": "same",
      "name": "Max",
      "price_usd_month": 168,
      "price_note": "$168 per AI Compare Ninja; earlier standard $160",
      "models": "GLM-5.3, GLM-5.3-Flash",
      "source": "https://www.aicompare.ninja/en/tool/zhipu-glm"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://docs.z.ai/devpack/overview"
     }
    ]
   },
   {
    "id": "kimi",
    "name": "Kimi Code (Moonshot)",
    "sources": [
     "https://www.kimi.ai/help/membership/membership-pricing",
     "https://www.kimi.ai/resources/kimi-code-introduction",
     "https://www.kimi.com/en/help/membership/membership-update-rules"
    ],
    "usage_model": "Kimi membership includes Kimi Code: rolling 5h limit (~300–1,200 requests), no weekly limit on new plans, shared monthly credit pool.",
    "plans": [
     {
      "window": "Rolling 5h (~300–1,200 requests)",
      "weekly_cap": "None on new plans; monthly shared credit pool",
      "overage": "Extra Usage (pay-as-you-go)",
      "coding_agent": "Kimi Code CLI + VS Code",
      "name": "Plus",
      "price_usd_month": 19,
      "models": "Kimi K2.7 Code; K3 on higher tiers",
      "source": "https://www.kimi.ai/help/membership/membership-pricing"
     },
     {
      "window": "Rolling 5h (~300–1,200 requests)",
      "weekly_cap": "None on new plans; monthly shared credit pool",
      "overage": "Extra Usage (pay-as-you-go)",
      "coding_agent": "Kimi Code CLI + VS Code",
      "name": "Pro",
      "price_usd_month": 39,
      "models": "Kimi K2.7 Code; K3 on higher tiers",
      "source": "https://www.kimi.ai/help/membership/membership-pricing"
     },
     {
      "window": "Rolling 5h (~300–1,200 requests)",
      "weekly_cap": "None on new plans; monthly shared credit pool",
      "overage": "Extra Usage (pay-as-you-go)",
      "coding_agent": "Kimi Code CLI + VS Code",
      "name": "Max",
      "price_usd_month": 99,
      "models": "Kimi K2.7 Code; K3 on higher tiers",
      "source": "https://www.kimi.ai/help/membership/membership-pricing"
     },
     {
      "window": "Rolling 5h (~300–1,200 requests)",
      "weekly_cap": "None on new plans; monthly shared credit pool",
      "overage": "Extra Usage (pay-as-you-go)",
      "coding_agent": "Kimi Code CLI + VS Code",
      "name": "Ultra",
      "price_usd_month": 199,
      "models": "Kimi K2.7 Code; K3 on higher tiers",
      "source": "https://www.kimi.ai/help/membership/membership-pricing"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://www.kimi.ai/help/membership/membership-pricing"
     }
    ]
   },
   {
    "id": "minimax",
    "name": "MiniMax Token Plan",
    "sources": [
     "https://platform.minimax.io/docs/guides/pricing-token-plan",
     "https://platform.minimax.io/docs/token-plan/faq"
    ],
    "usage_model": "Shared quota across text/image/speech with 5h rolling and weekly windows.",
    "plans": [
     {
      "window": "5h rolling (≈4,500 M2.7 calls)",
      "weekly_cap": "Weekly window (not published)",
      "overage": "Credits at pay-as-you-go ($1 = 1,000 credits)",
      "coding_agent": "MiniMax Code + Claude Code etc.",
      "name": "Plus",
      "price_usd_month": 22,
      "models": "MiniMax M3, M2.7",
      "source": "https://platform.minimax.io/docs/guides/pricing-token-plan"
     },
     {
      "window": "5h rolling (≈15,000 M2.7 calls)",
      "weekly_cap": "Weekly window",
      "overage": "Credits",
      "coding_agent": "same",
      "name": "Max",
      "price_usd_month": 55,
      "models": "MiniMax M3, M2.7",
      "source": "https://platform.minimax.io/docs/guides/pricing-token-plan"
     },
     {
      "window": "5h rolling (est. 15,000–36,000 M2.7 calls; no count from MiniMax)",
      "weekly_cap": "Weekly window",
      "overage": "Credits",
      "coding_agent": "same",
      "name": "Ultra",
      "price_usd_month": 132,
      "models": "MiniMax M3, M2.7",
      "source": "https://platform.minimax.io/docs/guides/pricing-token-plan"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://platform.minimax.io/docs/guides/pricing-token-plan"
     }
    ]
   },
   {
    "id": "alibaba",
    "name": "Alibaba Cloud Coding Plan (Qwen)",
    "sources": [
     "https://www.alibabacloud.com/help/en/model-studio/coding-plan",
     "https://docs.modelstudio.console.alibabacloud.com/en/model-studio/token-plan-overview"
    ],
    "usage_model": "Fixed monthly request quota; Lite discontinued; Pro sold in limited daily batches.",
    "plans": [
     {
      "window": "6,000 requests / 5h",
      "weekly_cap": "45,000 / week · 90,000 / month",
      "overage": "None",
      "coding_agent": "Qwen Code, Claude Code, Cline, OpenClaw",
      "name": "Coding Plan Pro",
      "price_usd_month": 50,
      "models": "Qwen3.7-Plus, Qwen3.6-Plus, Kimi K2.5, GLM-5, MiniMax-M2.5",
      "source": "https://www.alibabacloud.com/help/en/model-studio/coding-plan"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://www.alibabacloud.com/help/en/model-studio/coding-plan"
     }
    ]
   },
   {
    "id": "antigravity",
    "name": "Google Antigravity",
    "sources": [
     "https://antigravity.google/docs/plans/",
     "https://antigravity.google/docs/models",
     "https://antigravity.google/pricing/"
    ],
    "usage_model": "Via Google AI plans. Paid: quota refreshes every 5h up to a weekly cap; Free: weekly.",
    "plans": [
     {
      "window": "Weekly",
      "weekly_cap": "Weekly rate limit",
      "overage": "—",
      "coding_agent": "Antigravity IDE/CLI",
      "name": "Free",
      "price_usd_month": 0,
      "models": "Gemini 3.8 Flash, Claude 4.6 (until Nov 2), GPT-OSS",
      "source": "https://antigravity.google/docs/models"
     },
     {
      "window": "5h refresh",
      "weekly_cap": "Weekly cap",
      "overage": "AI credits",
      "coding_agent": "same",
      "name": "Google AI Pro",
      "price_usd_month": 19.99,
      "models": "+ Gemini 4 Argon, Claude Opus/Sonnet 5.5",
      "source": "https://antigravity.google/docs/models"
     },
     {
      "window": "5h refresh",
      "weekly_cap": "Highest weekly",
      "overage": "AI credits",
      "coding_agent": "same",
      "name": "AI Ultra 5x",
      "price_usd_month": 99.99,
      "models": "All incl. third-party",
      "source": "https://antigravity.google/pricing/"
     },
     {
      "window": "5h refresh",
      "weekly_cap": "Highest weekly",
      "overage": "AI credits",
      "coding_agent": "same",
      "name": "AI Ultra 20x",
      "price_usd_month": 199.99,
      "models": "All",
      "source": "https://antigravity.google/pricing/"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://antigravity.google/docs/plans/"
     }
    ]
   },
   {
    "id": "kiro",
    "name": "Kiro (AWS)",
    "sources": [
     "https://kiro.dev/pricing/",
     "https://kiro.dev/docs/billing/"
    ],
    "usage_model": "Monthly credits; add-ons $0.04/credit.",
    "plans": [
     {
      "window": "Monthly",
      "weekly_cap": "50 credits",
      "overage": "—",
      "coding_agent": "Kiro IDE/CLI/Web",
      "name": "Free",
      "price_usd_month": 0,
      "models": "Claude Sonnet 4.5 + open-weight",
      "source": "https://kiro.dev/pricing/"
     },
     {
      "window": "Monthly",
      "weekly_cap": "1,000 credits",
      "overage": "$0.04/credit add-ons",
      "coding_agent": "Kiro IDE/CLI/Web",
      "name": "Pro",
      "price_usd_month": 20,
      "models": "Auto, Claude Opus 5, Sonnet 5, GPT-5.6, open-weight",
      "source": "https://kiro.dev/pricing/"
     },
     {
      "window": "Monthly",
      "weekly_cap": "2,000 credits",
      "overage": "$0.04/credit add-ons",
      "coding_agent": "Kiro IDE/CLI/Web",
      "name": "Pro+",
      "price_usd_month": 40,
      "models": "Auto, Claude Opus 5, Sonnet 5, GPT-5.6, open-weight",
      "source": "https://kiro.dev/pricing/"
     },
     {
      "window": "Monthly",
      "weekly_cap": "5,000 credits",
      "overage": "$0.04/credit add-ons",
      "coding_agent": "Kiro IDE/CLI/Web",
      "name": "Pro Max",
      "price_usd_month": 100,
      "models": "Auto, Claude Opus 5, Sonnet 5, GPT-5.6, open-weight",
      "source": "https://kiro.dev/pricing/"
     },
     {
      "window": "Monthly",
      "weekly_cap": "10,000 credits",
      "overage": "$0.04/credit add-ons",
      "coding_agent": "Kiro IDE/CLI/Web",
      "name": "Power",
      "price_usd_month": 200,
      "models": "Auto, Claude Opus 5, Sonnet 5, GPT-5.6, open-weight",
      "source": "https://kiro.dev/pricing/"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://kiro.dev/pricing/"
     }
    ]
   },
   {
    "id": "jetbrains",
    "name": "JetBrains AI (Junie)",
    "sources": [
     "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html",
     "https://junie.jetbrains.com/docs/junie-cli-model-selection.html"
    ],
    "usage_model": "AI credits per 30 days; 1 credit = $1.",
    "plans": [
     {
      "window": "30 days",
      "weekly_cap": "3 AI credits",
      "overage": "—",
      "coding_agent": "Junie + AI Assistant",
      "name": "AI Free",
      "price_usd_month": 0,
      "models": "Junie Lite (Qwen 3.6 27B) + limited cloud",
      "source": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html"
     },
     {
      "window": "30 days",
      "weekly_cap": "10 AI credits",
      "overage": "Buy credits (12-month validity)",
      "coding_agent": "Junie + AI Assistant",
      "name": "AI Pro",
      "price_usd_month": 10,
      "models": "Claude Opus/Sonnet 5.5, GPT-6.1 Sol, Gemini, Grok 4.7",
      "source": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html"
     },
     {
      "window": "30 days",
      "weekly_cap": "35 AI credits",
      "overage": "Buy credits (12-month validity)",
      "coding_agent": "Junie + AI Assistant",
      "name": "AI Ultimate",
      "price_usd_month": 30,
      "models": "same",
      "source": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html"
     }
    ]
   },
   {
    "id": "warp",
    "name": "Warp",
    "sources": [
     "https://www.warp.dev/pricing"
    ],
    "usage_model": "$ of agent usage at API rates.",
    "plans": [
     {
      "window": "Monthly",
      "weekly_cap": "$20 agent usage",
      "overage": "Add-on usage",
      "coding_agent": "Warp Agent",
      "name": "Build",
      "price_usd_month": 20,
      "models": "Fable 5, Claude Opus 5, GPT-5.6, Kimi K3, GLM 5.2…",
      "source": "https://www.warp.dev/pricing"
     },
     {
      "window": "Monthly",
      "weekly_cap": "$200 agent usage",
      "overage": "Add-on usage",
      "coding_agent": "Warp Agent",
      "name": "Max",
      "price_usd_month": 200,
      "models": "same",
      "source": "https://www.warp.dev/pricing"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://www.warp.dev/pricing"
     }
    ]
   },
   {
    "id": "zed",
    "name": "Zed",
    "sources": [
     "https://zed.dev/docs/account/plans-and-pricing",
     "https://zed.dev/docs/account/zed-hosted-models.html"
    ],
    "usage_model": "$5 hosted credit, then provider price +10%.",
    "plans": [
     {
      "window": "Monthly",
      "weekly_cap": "$5 token credit",
      "overage": "List +10% (default $10 cap)",
      "coding_agent": "Zed Agent",
      "name": "Pro",
      "price_usd_month": 10,
      "models": "Claude Opus 5.5, Fable 5.1, GPT-6 Astra, Gemini…",
      "source": "https://zed.dev/docs/account/plans-and-pricing"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://zed.dev/docs/account/plans-and-pricing"
     }
    ]
   },
   {
    "id": "amp",
    "name": "Amp",
    "sources": [
     "https://ampcode.com/docs/pricing"
    ],
    "usage_model": "Monthly tier with included agent usage; overage at provider price, no markup.",
    "plans": [
     {
      "window": "—",
      "weekly_cap": "Pay as you go",
      "overage": "Credits at cost",
      "coding_agent": "Amp",
      "name": "Hobby",
      "price_usd_month": 0,
      "models": "Claude, GPT (BYOK/subscriptions)",
      "source": "https://ampcode.com/docs/pricing"
     },
     {
      "window": "Monthly",
      "weekly_cap": "Included agent usage + 45,000 orb-min",
      "overage": "Credits at cost",
      "coding_agent": "Amp",
      "name": "Individual",
      "price_usd_month": 20,
      "models": "Claude, GPT",
      "source": "https://ampcode.com/docs/pricing"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://ampcode.com/docs/pricing"
     }
    ]
   },
   {
    "id": "augment",
    "name": "Augment Code",
    "sources": [
     "https://www.augmentcode.com/pricing",
     "https://docs.augmentcode.com/models/token-based-pricing"
    ],
    "usage_model": "$ usage at API list price + 40% LLM fee.",
    "plans": [
     {
      "window": "Monthly",
      "weekly_cap": "$20 usage (≈$14.29 API tokens)",
      "overage": "Top-ups",
      "coding_agent": "Cosmos, Auggie CLI",
      "name": "Standard",
      "price_usd_month": 20,
      "models": "Claude Opus/Sonnet, GPT",
      "source": "https://www.augmentcode.com/pricing"
     },
     {
      "window": "Monthly",
      "weekly_cap": "$100 usage, up to 50 seats",
      "overage": "Top-ups",
      "coding_agent": "same",
      "name": "Business",
      "price_usd_month": 100,
      "models": "same",
      "source": "https://www.augmentcode.com/pricing"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://www.augmentcode.com/pricing"
     }
    ]
   },
   {
    "id": "factory",
    "name": "Factory (Droid)",
    "sources": [
     "https://docs.factory.ai/pricing/individuals",
     "https://docs.factory.ai/models"
    ],
    "usage_model": "Rolling 5h/7-day/30-day limits; Droid Core free pool afterwards.",
    "plans": [
     {
      "window": "5h / 7d / 30d rolling",
      "weekly_cap": "Standard Usage (≈20M tokens earlier)",
      "overage": "Extra Usage credits",
      "coding_agent": "Droid CLI/App",
      "name": "Pro",
      "price_usd_month": 20,
      "models": "All (Opus 5.5 at 1.6×, GPT-6 family)",
      "source": "https://docs.factory.ai/pricing/individuals"
     },
     {
      "window": "5h / 7d / 30d rolling",
      "weekly_cap": "~5x Pro",
      "overage": "Extra Usage credits",
      "coding_agent": "Droid CLI/App",
      "name": "Plus",
      "price_usd_month": 100,
      "models": "All (Opus 5.5 at 1.6×, GPT-6 family)",
      "source": "https://docs.factory.ai/pricing/individuals"
     },
     {
      "window": "5h / 7d / 30d rolling",
      "weekly_cap": "~10x Pro",
      "overage": "Extra Usage credits",
      "coding_agent": "Droid CLI/App",
      "name": "Max",
      "price_usd_month": 200,
      "models": "All (Opus 5.5 at 1.6×, GPT-6 family)",
      "source": "https://docs.factory.ai/pricing/individuals"
     }
    ],
    "changelog": [
     {
      "date": "2026-10-15",
      "change": "Added to PlanLimits.",
      "source": "https://docs.factory.ai/pricing/individuals"
     }
    ]
   }
  ]
 },
 "compare": {
  "checked": "2026-10-15",
  "benchmarks": {
   "tb4": {
    "name": "Terminal-Bench 4.0",
    "by": "Vals AI (same Mini-SWE-agent harness for every model)",
    "updated": "2026-10-13",
    "url": "https://www.vals.ai/benchmarks/terminal-bench-4",
    "scores": {
     "Claude Opus 5.5": 65.15,
     "Claude Sonnet 5.5": 64.14,
     "GPT-6 Astra": 59.6,
     "Claude Fable 5.1": 58.08,
     "Gemini 4 Argon": 57.58,
     "Claude Opus 5": 53.53,
     "Grok 4.7": 28.79,
     "Grok 4.6": 17.17,
     "GPT-6.1 Sol": 55.05,
     "GLM 5.3": 38.89,
     "Qwen 3.8 Max": 34.34,
     "Gemini 3.8 Flash": 19.7,
     "Kimi K3": 17.17,
     "GPT-6 Luna": 13.64,
     "Gemini 3.5 Flash": 6.06,
     "MiniMax-M3": 1.01,
     "Claude Opus 4.8": 23.23
    }
   },
   "aa": {
    "name": "Artificial Analysis Coding Agent Index v1.5",
    "by": "Artificial Analysis (avg of DeepSWE v1.1, Terminal-Bench 4.0, SWE-Atlas-QnA; scores are agent+model configs)",
    "updated": "2026-10-15 (snapshot via BenchLM mirror)",
    "url": "https://artificialanalysis.ai/agents/coding-agents",
    "mirror": "https://benchlm.ai/benchmarks/aacodingagents",
    "scores": {
     "Claude Code + Sonnet 5.5 (max)": 68.4,
     "Claude Code + Opus 5.5 (max)": 66,
     "Antigravity CLI + Gemini 4 Argon (high)": 63.8,
     "Codex + GPT-6 Astra (max)": 61.6,
     "Codex + GPT-6 Luna (max)": 41.1,
     "Grok Build + Grok 4.7 (xhigh)": 56.3,
     "Codex + GPT-6.1 Sol (xhigh)": 62.9,
     "Devin Fusion CLI + Opus 5.5 Max + GPT-6 Luna High": 68.1,
     "Claude Code + Opus 5 (max)": 59.7,
     "OpenCode + GLM-5.3 (max)": 53.6,
     "Kimi Code CLI + Kimi K3": 51.9,
     "Claude Code + Qwen3.8 Max": 43.3,
     "Antigravity SDK + Gemini 3.8 Flash (high)": 41.9
    }
   }
  },
  "tiers": [
   {
    "tier": "Free",
    "rows": [
     {
      "vendor": "Claude",
      "plan": "Free",
      "price": 0,
      "usage": "5-hour window, chat only (Anthropic gives no count)",
      "best_model": "Sonnet 5.5 (chat only)*",
      "tb4": 64.14,
      "note": "Chat only: Claude Code isn't included",
      "src": "https://claude.com/pricing",
      "value_id": "claude-free",
      "est": {},
      "aa_text": "n/a: no Claude Code"
     },
     {
      "vendor": "GitHub Copilot",
      "plan": "Free",
      "price": 0,
      "usage": "2,000 completions + 50 chat requests / mo",
      "best_model": "Claude Opus 5.5 (listed)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://github.com/features/copilot/plans",
      "value_id": "copilot-free",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Cursor",
      "plan": "Free",
      "price": 0,
      "usage": "Limited (not quantified)",
      "best_model": "Grok models only (Cursor)",
      "tb4": 28.79,
      "aa": 56.3,
      "aa_note": "Grok Build + Grok 4.7 (xhigh)",
      "src": "https://cursor.com/docs/models-and-pricing",
      "value_id": "free-na-0",
      "est": {
       "tb4": "Cursor says Free gets limited access to its first-party Grok models without naming versions. Shown: Grok 4.7, the best one, as an upper bound.",
       "aa": "Grok Build + Grok 4.7; same upper-bound caveat."
      }
     },
     {
      "vendor": "Windsurf",
      "plan": "Free",
      "price": 0,
      "usage": "Light daily+weekly quota (not quantified)",
      "best_model": "SWE-1.6 / SWE-2 (free in-house models)",
      "src": "https://docs.devin.ai/desktop/models",
      "value_id": "free-na-1",
      "est": {
       "best_model": "Free has 'limited model availability'. Free in-house models (SWE-1.5, SWE-2 promo through Oct 15) don't consume quota."
      },
      "tb4_text": "not benchmarked (est. below frontier)",
      "aa_text": "SWE-2 only as a Fusion sidekick"
     },
     {
      "vendor": "OpenAI Codex",
      "plan": "Free",
      "price": 0,
      "usage": "Quick tasks; est. at most Go's allowance",
      "best_model": "GPT-6 Luna (desktop app)",
      "tb4": 13.64,
      "aa": 41.1,
      "aa_note": "Codex + GPT-6 Luna (max)",
      "src": "https://developers.openai.com/codex/pricing/",
      "value_id": "free-na-2",
      "est": {
       "usage": "OpenAI publishes no Free allowance. Free is the cheapest plan, so we assume it gets no more than Go."
      }
     },
     {
      "vendor": "Google Antigravity",
      "plan": "Free",
      "price": 0,
      "usage": "Weekly quota (not quantified)",
      "best_model": "Gemini 3.8 Flash (+ Claude 4.6 until Nov 2)",
      "tb4": 19.7,
      "aa": 41.9,
      "aa_note": "Antigravity SDK + Gemini 3.8 Flash (high)",
      "src": "https://antigravity.google/docs/models",
      "value_id": "free-na-3",
      "est": {}
     },
     {
      "vendor": "Kiro",
      "plan": "Free",
      "price": 0,
      "usage": "50 credits / mo",
      "best_model": "Claude Sonnet 4.5 + open-weight models",
      "tb4": 23.23,
      "src": "https://kiro.dev/pricing/",
      "value_id": "kiro-free",
      "est": {
       "tb4": "Sonnet 4.5 isn't on Terminal-Bench 4.0. Est. ≈ Claude Opus 4.8 (23.2%), the closest older Claude on the board; probably generous for Sonnet 4.5.",
       "credits_usd": "Credits valued at Kiro's own $0.04 add-on price; a credit is a unit of work, not an API dollar."
      },
      "aa_text": "not on AA (est. low)",
      "credits_usd": 2
     },
     {
      "vendor": "JetBrains AI",
      "plan": "AI Free",
      "price": 0,
      "usage": "3 AI credits / 30 days (1 credit = $1)",
      "credits_usd": 3,
      "best_model": "Junie Lite: Qwen 3.6 27B",
      "src": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html",
      "value_id": "jb-free",
      "est": {},
      "tb4_text": "not benchmarked (small open model)",
      "aa_text": "not on AA"
     },
     {
      "vendor": "Amp",
      "plan": "Hobby",
      "price": 0,
      "usage": "No included agent usage; pay-as-you-go at provider price",
      "best_model": "Any (pay per use)",
      "tb4": 65.15,
      "src": "https://ampcode.com/docs/pricing",
      "value_id": "amp-hobby",
      "est": {
       "tb4": "Amp's model lineup isn't on its pricing page; it offers Claude and GPT models, so we show the best Claude score as an upper bound."
      },
      "aa_text": "Amp not on AA"
     }
    ]
   },
   {
    "tier": "~$8–10",
    "rows": [
     {
      "vendor": "Cursor",
      "plan": "Plus",
      "price": 8,
      "usage": "Base weekly pool (1x)",
      "best_model": "Grok 4.7 (pool also covers Grok 4.6, 4.5, Composer 2.5)",
      "tb4": 28.79,
      "aa": 56.3,
      "aa_note": "Grok Build + Grok 4.7 (xhigh)",
      "note": "Other providers' models: not available on Plus",
      "src": "https://cursor.com/docs/models-and-pricing",
      "value_id": "cursor-plus",
      "est": {
       "aa": "Cursor's own agent isn't on AA. Shown: the Grok Build + Grok 4.7 (xhigh) config. Plus runs medium effort, so real results may be lower."
      }
     },
     {
      "vendor": "GitHub Copilot",
      "plan": "Pro",
      "price": 10,
      "usage": "$15 AI credits / mo",
      "credits_usd": 15,
      "best_model": "Claude Opus 5.5 (listed)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://github.com/features/copilot/plans",
      "value_id": "copilot-pro",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "OpenAI Codex",
      "plan": "Go",
      "price": 8,
      "usage": "Est. ~140–3,000 GPT-6 Luna msgs / 5h (scaled from Plus)",
      "best_model": "GPT-6 Luna (desktop app)",
      "tb4": 13.64,
      "aa": 41.1,
      "aa_note": "Codex + GPT-6 Luna (max)",
      "src": "https://developers.openai.com/codex/pricing/",
      "value_id": "codex-go",
      "est": {
       "usage": "OpenAI publishes no Go numbers. Low = Plus's Luna floor (350) × 8/20 price ratio; high = Plus's Luna ceiling (3,000)."
      }
     },
     {
      "vendor": "JetBrains AI",
      "plan": "AI Pro",
      "price": 10,
      "usage": "10 AI credits / 30 days",
      "credits_usd": 10,
      "best_model": "Claude Opus 5.5 via Junie",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://junie.jetbrains.com/docs/junie-cli-model-selection.html",
      "value_id": "jb-pro",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Zed",
      "plan": "Pro",
      "price": 10,
      "usage": "$5 hosted-model credit (+ unlimited edit predictions); then list +10%",
      "credits_usd": 4.55,
      "best_model": "Claude Opus 5.5 / Fable 5.1 / GPT-6 Astra",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://zed.dev/docs/account/zed-hosted-models.html",
      "value_id": "zed-pro",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     }
    ]
   },
   {
    "tier": "~$20",
    "rows": [
     {
      "vendor": "Claude",
      "plan": "Pro",
      "price": 20,
      "usage": "5-hour window + weekly cap; est. 10–40 Claude Code prompts / 5h",
      "best_model": "Opus 5.5 / Sonnet 5.5*",
      "tb4": 65.15,
      "aa": 68.4,
      "aa_note": "Claude Code + Sonnet 5.5 (max)",
      "src": "https://claude.com/pricing",
      "value_id": "claude-pro",
      "est": {
       "usage": "Anthropic's last published figure (2025 help page, archived). 5-hour limits have been raised since, so this is likely a floor."
      }
     },
     {
      "vendor": "Windsurf",
      "plan": "Pro",
      "price": 20,
      "usage": "Daily+weekly quota; 7–27 Opus-class msgs/day (Mar 2026 table)",
      "best_model": "Claude Opus 5.5 via Fusion (lead model)",
      "tb4": 65.15,
      "aa": 68.1,
      "aa_note": "Devin Fusion CLI: Opus 5.5 Max + GPT-6 Luna",
      "src": "https://docs.devin.ai/desktop/models",
      "value_id": "windsurf-pro",
      "est": {
       "best_model": "Windsurf/Devin Desktop has 'full model availability' on paid plans and recommends Fusion. The AA entry for Fusion uses Opus 5.5 as the lead, so we assume Opus 5.5 is included.",
       "tb4": "Opus 5.5's own score (assumed included).",
       "aa": "Devin Fusion CLI (same vendor, CLI rather than Desktop)."
      }
     },
     {
      "vendor": "Gemini Code Assist",
      "plan": "Standard",
      "price": 22.8,
      "usage": "1,500 agent/CLI requests / day",
      "best_model": "Gemini 3.5 Flash (default; agent mode picks the model)",
      "tb4": 6.06,
      "aa": 41.9,
      "aa_note": "est.: Antigravity SDK + Gemini 3.8 Flash",
      "note": "No new purchases since 2026-10-09",
      "src": "https://developers.google.com/gemini-code-assist/docs/gemini-3",
      "value_id": "gemini-std",
      "est": {
       "aa": "Gemini 3.5 Flash isn't on AA. Shown: the newer Gemini 3.8 Flash config, an upper bound."
      }
     },
     {
      "vendor": "OpenAI Codex",
      "plan": "Plus",
      "price": 20,
      "usage": "Per 5h (OpenAI est.): 5–45 GPT-6 Astra, 15–160 GPT-6.1 Sol or 350–3,000 GPT-6 Luna; weekly limits may apply",
      "best_model": "GPT-6 Astra / GPT-6.1 Sol",
      "tb4": 59.6,
      "aa": 62.9,
      "aa_note": "Codex + GPT-6.1 Sol (xhigh)",
      "src": "https://developers.openai.com/codex/pricing/",
      "value_id": "codex-plus",
      "est": {}
     },
     {
      "vendor": "Kimi Code",
      "plan": "Plus",
      "price": 19,
      "usage": "Rolling 5h window, ~300–1,200 requests (Kimi, not split by plan); no weekly cap on new plans; shared monthly credit pool",
      "best_model": "Kimi K2.7 Code (default); K3 on higher tiers*",
      "tb4": 17.17,
      "aa": 51.9,
      "aa_note": "Kimi Code CLI + Kimi K3",
      "src": "https://www.kimi.ai/help/membership/membership-pricing",
      "value_id": "kimi-plus",
      "est": {
       "tb4": "Kimi K3's score. K2.7 Code (the default) isn't on Terminal-Bench 4.0, so this is an upper bound. Which new tiers unlock K3 isn't stated (the docs use legacy plan names).",
       "aa": "Kimi Code CLI + K3, the vendor's own agent. Same caveat: K3 tier mapping isn't stated."
      }
     },
     {
      "vendor": "Z.ai GLM",
      "plan": "Coding Lite",
      "price": 18,
      "usage": "2,000 credits / 5h · 10,000 / week (≈48–97M GLM-5.3 tokens/week)",
      "best_model": "GLM-5.3 (+ GLM-5.3-Flash)",
      "tb4": 38.89,
      "aa": 53.6,
      "aa_note": "OpenCode + GLM-5.3 (max)",
      "note": "Works in Claude Code, OpenCode, Cline…",
      "src": "https://docs.z.ai/devpack/overview",
      "value_id": "zai-lite",
      "est": {}
     },
     {
      "vendor": "MiniMax",
      "plan": "Token Plus",
      "price": 22,
      "usage": "5h + weekly windows; 4,500 M2.7 calls / 5h (third-party capture)",
      "best_model": "MiniMax M3 / M2.7",
      "tb4": 1.01,
      "src": "https://platform.minimax.io/docs/guides/pricing-token-plan",
      "value_id": "minimax-plus",
      "est": {
       "aa": "No MiniMax model is on the AA index. MiniMax-M3 scores 1.0% on Terminal-Bench 4.0, so we expect it to rank near the bottom.",
       "usage": "The call count comes from a Sept 23 capture of MiniMax's subscribe page by iadecider.com; MiniMax's docs now list only the windows."
      },
      "aa_text": "not on AA (est. low)"
     },
     {
      "vendor": "Google Antigravity",
      "plan": "Google AI Pro",
      "price": 19.99,
      "usage": "5h refresh + weekly cap (not quantified)",
      "best_model": "Claude Opus 5.5 / Gemini 4 Argon",
      "tb4": 65.15,
      "aa": 63.8,
      "aa_note": "Antigravity CLI + Gemini 4 Argon (high)",
      "note": "Via Google AI plans; quota refreshes every 5h up to a weekly cap",
      "src": "https://antigravity.google/docs/models",
      "value_id": "ag-pro",
      "est": {}
     },
     {
      "vendor": "Kiro",
      "plan": "Pro",
      "price": 20,
      "usage": "1,000 credits / mo (add-ons $0.04 each)",
      "best_model": "Claude Opus 5 (listed for paid plans)",
      "tb4": 53.53,
      "aa": 59.7,
      "aa_note": "est. via Claude Code + Opus 5",
      "src": "https://kiro.dev/pricing/",
      "value_id": "kiro-pro",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair.",
       "credits_usd": "Credits valued at Kiro's own $0.04 add-on price; a credit is a unit of work, not an API dollar."
      },
      "credits_usd": 40
     },
     {
      "vendor": "Warp",
      "plan": "Build",
      "price": 20,
      "usage": "$20 agent usage / mo at API rates",
      "credits_usd": 20,
      "best_model": "Claude Opus 5 (listed)",
      "tb4": 53.53,
      "aa": 59.7,
      "aa_note": "est. via Claude Code + Opus 5",
      "src": "https://www.warp.dev/pricing",
      "value_id": "warp-build",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Amp",
      "plan": "Individual",
      "price": 20,
      "usage": "Included agent usage (≥$20 per trackers) + 45,000 orb-minutes",
      "credits_usd": 20,
      "best_model": "Claude / GPT frontier (versions not on pricing page)",
      "tb4": 65.15,
      "src": "https://ampcode.com/docs/pricing",
      "value_id": "amp-ind",
      "est": {
       "credits_usd": "Amp doesn't print a $ figure; ≥$20 per third-party trackers.",
       "tb4": "Best Claude score, assuming Amp offers Opus 5.5 (it supports Claude subscriptions and API)."
      },
      "aa_text": "Amp not on AA"
     },
     {
      "vendor": "Augment Code",
      "plan": "Standard",
      "price": 20,
      "usage": "$20 usage incl. 40% LLM fee (≈$14 of API tokens); up to 50 seats",
      "credits_usd": 14.29,
      "best_model": "Claude Opus / Sonnet (per Augment docs)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://docs.augmentcode.com/models/token-based-pricing",
      "value_id": "aug-std",
      "est": {
       "tb4": "Augment's docs show Opus and Sonnet tasks; we assume the current Opus 5.5.",
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Factory",
      "plan": "Pro",
      "price": 20,
      "usage": "Rolling 5h / 7-day / 30-day limits (≈20M Standard Tokens/mo earlier)",
      "best_model": "Claude Opus 5.5 / GPT-6 Astra (all models on Standard Usage)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://docs.factory.ai/models",
      "value_id": "factory-pro",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     }
    ]
   },
   {
    "tier": "~$30–40",
    "rows": [
     {
      "vendor": "Cursor",
      "plan": "Premium",
      "price": 30,
      "usage": "4x Plus weekly pool",
      "best_model": "Grok 4.7 (pool also covers Grok 4.6, 4.5, Composer 2.5)",
      "tb4": 28.79,
      "aa": 56.3,
      "aa_note": "Grok Build + Grok 4.7 (xhigh)",
      "note": "Other providers' models: on-demand at API price",
      "src": "https://cursor.com/docs/models-and-pricing",
      "value_id": "cursor-premium",
      "est": {
       "aa": "Cursor's own agent isn't on AA. Shown: the Grok Build + Grok 4.7 (xhigh) config. Plus runs medium effort, so real results may be lower."
      }
     },
     {
      "vendor": "GitHub Copilot",
      "plan": "Pro+",
      "price": 39,
      "usage": "$70 AI credits / mo",
      "credits_usd": 70,
      "best_model": "Claude Opus 5.5 (listed)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://github.com/features/copilot/plans",
      "value_id": "copilot-proplus",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Windsurf",
      "plan": "Teams (per seat)",
      "price": 40,
      "usage": "Pro-level quota per seat",
      "best_model": "Claude Opus 5.5 via Fusion (lead model)",
      "tb4": 65.15,
      "aa": 68.1,
      "aa_note": "Devin Fusion CLI: Opus 5.5 Max + GPT-6 Luna",
      "src": "https://docs.devin.ai/desktop/models",
      "value_id": "windsurf-teams",
      "est": {
       "best_model": "Windsurf/Devin Desktop has 'full model availability' on paid plans and recommends Fusion. The AA entry for Fusion uses Opus 5.5 as the lead, so we assume Opus 5.5 is included.",
       "tb4": "Opus 5.5's own score (assumed included).",
       "aa": "Devin Fusion CLI (same vendor, CLI rather than Desktop)."
      }
     },
     {
      "vendor": "JetBrains AI",
      "plan": "AI Ultimate",
      "price": 30,
      "usage": "35 AI credits / 30 days",
      "credits_usd": 35,
      "best_model": "Claude Opus 5.5 via Junie",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://junie.jetbrains.com/docs/junie-cli-model-selection.html",
      "value_id": "jb-ult",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Kimi Code",
      "plan": "Pro",
      "price": 39,
      "usage": "Rolling 5h window, ~300–1,200 requests (Kimi, not split by plan); no weekly cap on new plans; shared monthly credit pool",
      "best_model": "Kimi K2.7 Code (default); K3 on higher tiers*",
      "tb4": 17.17,
      "aa": 51.9,
      "aa_note": "Kimi Code CLI + Kimi K3",
      "src": "https://www.kimi.ai/help/membership/membership-pricing",
      "value_id": "kimi-pro",
      "est": {
       "tb4": "Kimi K3's score. K2.7 Code (the default) isn't on Terminal-Bench 4.0, so this is an upper bound. Which new tiers unlock K3 isn't stated (the docs use legacy plan names).",
       "aa": "Kimi Code CLI + K3, the vendor's own agent. Same caveat: K3 tier mapping isn't stated."
      }
     },
     {
      "vendor": "Kiro",
      "plan": "Pro+",
      "price": 40,
      "usage": "2,000 credits / mo (add-ons $0.04 each)",
      "best_model": "Claude Opus 5 (listed for paid plans)",
      "tb4": 53.53,
      "aa": 59.7,
      "aa_note": "est. via Claude Code + Opus 5",
      "src": "https://kiro.dev/pricing/",
      "value_id": "kiro-proplus",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair.",
       "credits_usd": "Credits valued at Kiro's own $0.04 add-on price; a credit is a unit of work, not an API dollar."
      },
      "credits_usd": 80
     }
    ]
   },
   {
    "tier": "~$50–80",
    "rows": [
     {
      "vendor": "Alibaba Model Studio",
      "plan": "Coding Plan Pro",
      "price": 50,
      "usage": "6,000 requests / 5h · 45,000 / week · 90,000 / month",
      "best_model": "Qwen3.7-Plus (+ Kimi K2.5, GLM-5, MiniMax M2.5)",
      "tb4": 34.34,
      "aa": 43.3,
      "aa_note": "est.: Claude Code + Qwen3.8 Max",
      "note": "Limited stock; restocked daily",
      "src": "https://www.alibabacloud.com/help/en/model-studio/coding-plan",
      "value_id": "ali-pro",
      "est": {
       "tb4": "Qwen3.7-Plus isn't on Terminal-Bench 4.0. Shown: the newer Qwen 3.8 Max (34.3%), so treat it as an upper bound.",
       "aa": "Qwen3.8 Max in Claude Code; same upper-bound caveat."
      }
     },
     {
      "vendor": "MiniMax",
      "plan": "Token Max",
      "price": 55,
      "usage": "5h + weekly windows; 15,000 M2.7 calls / 5h (third-party capture)",
      "best_model": "MiniMax M3 / M2.7",
      "tb4": 1.01,
      "src": "https://platform.minimax.io/docs/guides/pricing-token-plan",
      "value_id": "minimax-max",
      "est": {
       "aa": "No MiniMax model is on the AA index. MiniMax-M3 scores 1.0% on Terminal-Bench 4.0, so we expect it to rank near the bottom.",
       "usage": "From the same Sept 23 third-party capture."
      },
      "aa_text": "not on AA (est. low)"
     },
     {
      "vendor": "Z.ai GLM",
      "plan": "Coding Pro",
      "price": 80,
      "usage": "12,000 credits / 5h · 60,000 / week (≈290–580M tokens/week)",
      "best_model": "GLM-5.3 (+ GLM-5.3-Flash)",
      "tb4": 38.89,
      "aa": 53.6,
      "aa_note": "OpenCode + GLM-5.3 (max)",
      "note": "Works in Claude Code, OpenCode, Cline…",
      "src": "https://docs.z.ai/devpack/overview",
      "value_id": "zai-pro",
      "est": {
       "price": "Z.ai docs say 'from $18'. Pro/Max list prices from AI Compare Ninja (Oct 2026); Z.ai's migration notice shows $72/$160 as the earlier standard prices."
      }
     }
    ]
   },
   {
    "tier": "~$100",
    "rows": [
     {
      "vendor": "Claude",
      "plan": "Max 5x",
      "price": 100,
      "usage": "5x Pro per session + weekly cap",
      "best_model": "Opus 5.5 / Sonnet 5.5*",
      "tb4": 65.15,
      "aa": 68.4,
      "aa_note": "Claude Code + Sonnet 5.5 (max)",
      "src": "https://claude.com/pricing",
      "value_id": "claude-max5",
      "est": {}
     },
     {
      "vendor": "OpenAI Codex",
      "plan": "Pro 100",
      "price": 100,
      "usage": "No 5-hour limit; weekly allowance = 5x Plus (OpenAI on X)",
      "best_model": "GPT-6 Astra / GPT-6.1 Sol",
      "tb4": 59.6,
      "aa": 62.9,
      "aa_note": "Codex + GPT-6.1 Sol (xhigh)",
      "src": "https://developers.openai.com/codex/pricing/",
      "value_id": "codex-pro100",
      "est": {}
     },
     {
      "vendor": "Cursor",
      "plan": "Super",
      "price": 100,
      "usage": "15x Plus weekly pool",
      "best_model": "Grok 4.7 (pool also covers Grok 4.6, 4.5, Composer 2.5)",
      "tb4": 28.79,
      "aa": 56.3,
      "aa_note": "Grok Build + Grok 4.7 (xhigh)",
      "note": "Other providers' models: on-demand at API price",
      "src": "https://cursor.com/docs/models-and-pricing",
      "value_id": "cursor-super",
      "est": {
       "aa": "Cursor's own agent isn't on AA. Shown: the Grok Build + Grok 4.7 (xhigh) config. Plus runs medium effort, so real results may be lower."
      }
     },
     {
      "vendor": "GitHub Copilot",
      "plan": "Max",
      "price": 100,
      "usage": "$200 AI credits / mo",
      "credits_usd": 200,
      "best_model": "Claude Opus 5.5 (listed)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://github.com/features/copilot/plans",
      "value_id": "copilot-max",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Kimi Code",
      "plan": "Max",
      "price": 99,
      "usage": "Rolling 5h window, ~300–1,200 requests (Kimi, not split by plan); no weekly cap on new plans; shared monthly credit pool",
      "best_model": "Kimi K2.7 Code (default); K3 on higher tiers*",
      "tb4": 17.17,
      "aa": 51.9,
      "aa_note": "Kimi Code CLI + Kimi K3",
      "src": "https://www.kimi.ai/help/membership/membership-pricing",
      "value_id": "kimi-max",
      "est": {
       "tb4": "Kimi K3's score. K2.7 Code (the default) isn't on Terminal-Bench 4.0, so this is an upper bound. Which new tiers unlock K3 isn't stated (the docs use legacy plan names).",
       "aa": "Kimi Code CLI + K3, the vendor's own agent. Same caveat: K3 tier mapping isn't stated."
      }
     },
     {
      "vendor": "Google Antigravity",
      "plan": "AI Ultra 5x",
      "price": 99.99,
      "usage": "5x Pro allowance; highest weekly limits",
      "best_model": "Claude Opus 5.5 / Gemini 4 Argon",
      "tb4": 65.15,
      "aa": 63.8,
      "aa_note": "Antigravity CLI + Gemini 4 Argon (high)",
      "note": "Via Google AI plans; quota refreshes every 5h up to a weekly cap",
      "src": "https://antigravity.google/docs/models",
      "value_id": "ag-ultra5",
      "est": {}
     },
     {
      "vendor": "Kiro",
      "plan": "Pro Max",
      "price": 100,
      "usage": "5,000 credits / mo (add-ons $0.04 each)",
      "best_model": "Claude Opus 5 (listed for paid plans)",
      "tb4": 53.53,
      "aa": 59.7,
      "aa_note": "est. via Claude Code + Opus 5",
      "src": "https://kiro.dev/pricing/",
      "value_id": "kiro-promax",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair.",
       "credits_usd": "Credits valued at Kiro's own $0.04 add-on price; a credit is a unit of work, not an API dollar."
      },
      "credits_usd": 200
     },
     {
      "vendor": "Augment Code",
      "plan": "Business (team)",
      "price": 100,
      "usage": "$100 usage incl. 40% LLM fee; up to 50 seats",
      "credits_usd": 71.43,
      "best_model": "Claude Opus / Sonnet (per Augment docs)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://docs.augmentcode.com/models/token-based-pricing",
      "value_id": "aug-biz",
      "est": {
       "tb4": "Assumes Opus 5.5 (see Standard).",
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Factory",
      "plan": "Plus",
      "price": 100,
      "usage": "~5x Pro usage",
      "best_model": "Claude Opus 5.5 / GPT-6 Astra (all models on Standard Usage)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://docs.factory.ai/models",
      "value_id": "factory-plus",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     }
    ]
   },
   {
    "tier": "~$130–200",
    "rows": [
     {
      "vendor": "Claude",
      "plan": "Max 20x",
      "price": 200,
      "usage": "20x Pro per session + weekly cap",
      "best_model": "Opus 5.5 / Sonnet 5.5*",
      "tb4": 65.15,
      "aa": 68.4,
      "aa_note": "Claude Code + Sonnet 5.5 (max)",
      "src": "https://claude.com/pricing",
      "value_id": "claude-max20",
      "est": {}
     },
     {
      "vendor": "OpenAI Codex",
      "plan": "Pro 200",
      "price": 200,
      "usage": "No 5-hour limit; 10x Plus for new subs (20x for existing until Oct 29)",
      "best_model": "GPT-6 Astra / GPT-6.1 Sol",
      "tb4": 59.6,
      "aa": 62.9,
      "aa_note": "Codex + GPT-6.1 Sol (xhigh)",
      "src": "https://developers.openai.com/codex/pricing/",
      "value_id": "codex-pro200",
      "est": {}
     },
     {
      "vendor": "Cursor",
      "plan": "Ultra",
      "price": 200,
      "usage": "50x Plus weekly pool",
      "best_model": "Grok 4.7 (pool also covers Grok 4.6, 4.5, Composer 2.5)",
      "tb4": 28.79,
      "aa": 56.3,
      "aa_note": "Grok Build + Grok 4.7 (xhigh)",
      "note": "Other providers' models: on-demand at API price",
      "src": "https://cursor.com/docs/models-and-pricing",
      "value_id": "cursor-ultra",
      "est": {
       "aa": "Cursor's own agent isn't on AA. Shown: the Grok Build + Grok 4.7 (xhigh) config. Plus runs medium effort, so real results may be lower."
      }
     },
     {
      "vendor": "Windsurf",
      "plan": "Max",
      "price": 200,
      "usage": "42–170 Opus-class msgs/day (Mar 2026 table)",
      "best_model": "Claude Opus 5.5 via Fusion (lead model)",
      "tb4": 65.15,
      "aa": 68.1,
      "aa_note": "Devin Fusion CLI: Opus 5.5 Max + GPT-6 Luna",
      "src": "https://docs.devin.ai/desktop/models",
      "value_id": "windsurf-max",
      "est": {
       "best_model": "Windsurf/Devin Desktop has 'full model availability' on paid plans and recommends Fusion. The AA entry for Fusion uses Opus 5.5 as the lead, so we assume Opus 5.5 is included.",
       "tb4": "Opus 5.5's own score (assumed included).",
       "aa": "Devin Fusion CLI (same vendor, CLI rather than Desktop)."
      }
     },
     {
      "vendor": "Kimi Code",
      "plan": "Ultra",
      "price": 199,
      "usage": "Rolling 5h window, ~300–1,200 requests (Kimi, not split by plan); no weekly cap on new plans; shared monthly credit pool",
      "best_model": "Kimi K2.7 Code (default); K3 on higher tiers*",
      "tb4": 17.17,
      "aa": 51.9,
      "aa_note": "Kimi Code CLI + Kimi K3",
      "src": "https://www.kimi.ai/help/membership/membership-pricing",
      "value_id": "kimi-ultra",
      "est": {
       "tb4": "Kimi K3's score. K2.7 Code (the default) isn't on Terminal-Bench 4.0, so this is an upper bound. Which new tiers unlock K3 isn't stated (the docs use legacy plan names).",
       "aa": "Kimi Code CLI + K3, the vendor's own agent. Same caveat: K3 tier mapping isn't stated."
      }
     },
     {
      "vendor": "Z.ai GLM",
      "plan": "Coding Max",
      "price": 168,
      "usage": "28,000 credits / 5h · 140,000 / week",
      "best_model": "GLM-5.3 (+ GLM-5.3-Flash)",
      "tb4": 38.89,
      "aa": 53.6,
      "aa_note": "OpenCode + GLM-5.3 (max)",
      "note": "Works in Claude Code, OpenCode, Cline…",
      "src": "https://docs.z.ai/devpack/overview",
      "value_id": "zai-max",
      "est": {
       "price": "Z.ai docs say 'from $18'. Pro/Max list prices from AI Compare Ninja (Oct 2026); Z.ai's migration notice shows $72/$160 as the earlier standard prices."
      }
     },
     {
      "vendor": "MiniMax",
      "plan": "Token Ultra",
      "price": 132,
      "usage": "5h + weekly windows; est. 15,000–36,000 M2.7 calls / 5h",
      "best_model": "MiniMax M3 / M2.7",
      "tb4": 1.01,
      "src": "https://platform.minimax.io/docs/guides/pricing-token-plan",
      "value_id": "minimax-ultra",
      "est": {
       "aa": "No MiniMax model is on the AA index. MiniMax-M3 scores 1.0% on Terminal-Bench 4.0, so we expect it to rank near the bottom.",
       "usage": "MiniMax publishes no call count for Ultra; the Real value estimate is bounded by Max and the price ratio."
      },
      "aa_text": "not on AA (est. low)"
     },
     {
      "vendor": "Google Antigravity",
      "plan": "AI Ultra 20x",
      "price": 199.99,
      "usage": "20x Pro allowance",
      "best_model": "Claude Opus 5.5 / Gemini 4 Argon",
      "tb4": 65.15,
      "aa": 63.8,
      "aa_note": "Antigravity CLI + Gemini 4 Argon (high)",
      "note": "Via Google AI plans; quota refreshes every 5h up to a weekly cap",
      "src": "https://antigravity.google/docs/models",
      "value_id": "ag-ultra20",
      "est": {}
     },
     {
      "vendor": "Kiro",
      "plan": "Power",
      "price": 200,
      "usage": "10,000 credits / mo (add-ons $0.04 each)",
      "best_model": "Claude Opus 5 (listed for paid plans)",
      "tb4": 53.53,
      "aa": 59.7,
      "aa_note": "est. via Claude Code + Opus 5",
      "src": "https://kiro.dev/pricing/",
      "value_id": "kiro-power",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair.",
       "credits_usd": "Credits valued at Kiro's own $0.04 add-on price; a credit is a unit of work, not an API dollar."
      },
      "credits_usd": 400
     },
     {
      "vendor": "Warp",
      "plan": "Max",
      "price": 200,
      "usage": "$200 agent usage / mo (10x Build)",
      "credits_usd": 200,
      "best_model": "Claude Opus 5 (listed)",
      "tb4": 53.53,
      "aa": 59.7,
      "aa_note": "est. via Claude Code + Opus 5",
      "src": "https://www.warp.dev/pricing",
      "value_id": "warp-max",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     },
     {
      "vendor": "Factory",
      "plan": "Max",
      "price": 200,
      "usage": "~10x Pro usage",
      "best_model": "Claude Opus 5.5 / GPT-6 Astra (all models on Standard Usage)",
      "tb4": 65.15,
      "aa": 66,
      "aa_note": "est. via Claude Code + Opus 5.5",
      "src": "https://docs.factory.ai/models",
      "value_id": "factory-max",
      "est": {
       "aa": "Harness differs: this is the best AA score for the same model in another agent (Claude Code). Products weren't benchmarked as a pair."
      }
     }
    ]
   }
  ]
 },
 "value": {
  "checked": "2026-10-15",
  "prices": {
   "Claude Opus 5.5": {
    "in": 4,
    "cached": 0.2,
    "out": 20,
    "src": "https://platform.claude.com/docs/en/about-claude/pricing"
   },
   "Claude Sonnet 5.5": {
    "in": 2,
    "cached": 0.1,
    "out": 10,
    "src": "https://platform.claude.com/docs/en/about-claude/pricing"
   },
   "Claude Fable 5.1": {
    "in": 10,
    "cached": 0.25,
    "out": 50,
    "src": "https://platform.claude.com/docs/en/about-claude/pricing"
   },
   "Claude Haiku 5.5": {
    "in": 0.1,
    "cached": 0.01,
    "out": 0.5,
    "note": "prompts ≤100K tokens; $0.50/$0.05/$2.50 above",
    "src": "https://platform.claude.com/docs/en/about-claude/pricing"
   },
   "GPT-6 Astra": {
    "in": 10,
    "cached": 1,
    "out": 50,
    "note": "≤272K context",
    "src": "https://developers.openai.com/api/docs/pricing"
   },
   "GPT-6.1 Sol": {
    "in": 2,
    "cached": 0.1,
    "out": 10,
    "src": "https://developers.openai.com/api/docs/pricing"
   },
   "GPT-6 Luna": {
    "in": 0.1,
    "cached": 0.01,
    "out": 0.5,
    "src": "https://developers.openai.com/api/docs/pricing"
   },
   "GPT-5.6 Luna": {
    "in": 0.2,
    "cached": 0.02,
    "out": 1.2,
    "src": "https://developers.openai.com/api/docs/pricing"
   },
   "Grok 4.7": {
    "in": 2,
    "cached": 0.5,
    "out": 6,
    "note": "Cursor's published rate beyond the pool",
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   "Grok 4.6": {
    "in": 2,
    "cached": 0.5,
    "out": 6,
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   "Grok 4.5": {
    "in": 2,
    "cached": 0.5,
    "out": 6,
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   "Composer 2.5": {
    "in": 0.5,
    "cached": 0.2,
    "out": 2.5,
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   "Gemini 4 Argon": {
    "in": 2,
    "cached": 0.1,
    "out": 10,
    "note": "as listed by Cursor at API list price",
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   "Claude Opus 5": {
    "in": 5,
    "cached": 0.5,
    "out": 25,
    "note": "provider price as listed by Zed",
    "src": "https://zed.dev/docs/account/zed-hosted-models.html"
   },
   "Claude Opus 4.6": {
    "in": 5,
    "cached": 0.5,
    "out": 25,
    "note": "provider price as listed by Zed",
    "src": "https://zed.dev/docs/account/zed-hosted-models.html"
   },
   "Claude Sonnet 4.5": {
    "in": 3,
    "cached": 0.3,
    "out": 15,
    "note": "provider price as listed by Zed",
    "src": "https://zed.dev/docs/account/zed-hosted-models.html"
   },
   "GLM-5.3": {
    "in": 1.4,
    "cached": 0.345,
    "out": 4.4,
    "note": "in/out from Vals AI; cached est. = input × Z.ai's cached/input credit ratio (1.7/6.9)",
    "est": true,
    "src": "https://www.vals.ai/benchmarks/terminal-bench-4"
   },
   "Kimi K2.7 Code": {
    "in": 0.95,
    "cached": 0.19,
    "out": 4,
    "src": "https://platform.kimi.ai/"
   },
   "MiniMax M2.7": {
    "in": 0.3,
    "cached": 0.06,
    "out": 1.2,
    "src": "https://platform.minimax.io/docs/pricing/overview"
   },
   "Qwen3.7-Plus": {
    "in": 0.4,
    "cached": 0.04,
    "out": 1.6,
    "note": "≤256K tokens; cached est. at 10% of input (not on the price page we read)",
    "est": true,
    "src": "https://www.alibabacloud.com/help/en/model-studio/model-pricing"
   },
   "Gemini 3.5 Flash": {
    "in": 1.5,
    "cached": 0.15,
    "out": 9,
    "note": "provider price as listed by Zed; cached est. at 10% of input",
    "est": true,
    "src": "https://zed.dev/docs/account/zed-hosted-models.html"
   },
   "SWE-2": {
    "in": 3,
    "cached": 0.3,
    "out": 15,
    "note": "Cognition list price (75% off through Dec 31, 2026)",
    "src": "https://docs.devin.ai/desktop/models"
   }
  },
  "per_call": {
   "cached": 65300,
   "fresh": 12500,
   "out": 480,
   "derivation": "BSWEN: 1,289 Claude Code requests = 100.3M input (84.2M cache reads) + 616K output → per request ≈ 65.3K cached, 12.5K fresh, 0.48K output.",
   "src": "https://docs.bswen.com/blog/2026-03-10-claude-code-token-usage-per-request/"
  },
  "profiles": {
   "light": {
    "label": "Light",
    "calls": 5,
    "out_mult": 1,
    "blurb": "Quick fix: a few file reads, one edit."
   },
   "typical": {
    "label": "Typical",
    "calls": 15,
    "out_mult": 3,
    "blurb": "A normal agent turn: explore, edit, run tests, fix."
   },
   "heavy": {
    "label": "Heavy",
    "calls": 40,
    "out_mult": 10,
    "blurb": "Long autonomous run on a big repo with lots of reasoning."
   }
  },
  "default_windows": 20,
  "plans": [
   {
    "id": "claude-free",
    "vendor": "Claude",
    "plan": "Free",
    "price": 0,
    "kind": "zero",
    "official": true,
    "caveat": "Chat only: Claude Code isn't on Free.",
    "src": "https://claude.com/pricing"
   },
   {
    "id": "copilot-free",
    "vendor": "GitHub Copilot",
    "plan": "Free",
    "price": 0,
    "kind": "window",
    "unit": "month",
    "official": false,
    "models": [
     {
      "m": "Claude Sonnet 5.5",
      "lo": 50,
      "hi": 50
     }
    ],
    "caveat": "50 chat requests a month (GitHub). Est.: priced as one Typical Sonnet 5.5 turn each; GitHub doesn't say which models Free requests reach.",
    "src": "https://github.com/features/copilot/plans"
   },
   {
    "id": "kiro-free",
    "vendor": "Kiro",
    "plan": "Free",
    "price": 0,
    "kind": "credits",
    "credits": 2,
    "official": false,
    "caveat": "50 credits × Kiro's $0.04 add-on credit price. A credit is Kiro's unit of work, not an API dollar.",
    "src": "https://kiro.dev/docs/billing/"
   },
   {
    "id": "jb-free",
    "vendor": "JetBrains AI",
    "plan": "AI Free",
    "price": 0,
    "kind": "credits",
    "credits": 3,
    "official": true,
    "caveat": "3 AI credits per 30 days; JetBrains: 1 credit = $1.",
    "src": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html"
   },
   {
    "id": "amp-hobby",
    "vendor": "Amp",
    "plan": "Hobby",
    "price": 0,
    "kind": "zero",
    "official": true,
    "caveat": "No included agent usage; pay-as-you-go at provider prices (no markup).",
    "src": "https://ampcode.com/docs/pricing"
   },
   {
    "id": "free-na-0",
    "vendor": "Cursor",
    "plan": "Free",
    "price": 0,
    "kind": "na",
    "official": false,
    "why": "Limited Grok-only quota, not quantified by Cursor.",
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   {
    "id": "free-na-1",
    "vendor": "Windsurf",
    "plan": "Free",
    "price": 0,
    "kind": "na",
    "official": false,
    "why": "Light daily/weekly quota, not quantified; free models (e.g. SWE-1.5) don't count against it.",
    "src": "https://devin.ai/blog/windsurf-pricing-plans"
   },
   {
    "id": "free-na-2",
    "vendor": "OpenAI Codex",
    "plan": "Free",
    "price": 0,
    "kind": "na",
    "official": false,
    "why": "GPT-6 Luna in the desktop app; OpenAI publishes no Free allowance (assume ≤ Go).",
    "src": "https://developers.openai.com/codex/pricing/"
   },
   {
    "id": "free-na-3",
    "vendor": "Google Antigravity",
    "plan": "Free",
    "price": 0,
    "kind": "na",
    "official": false,
    "why": "'A meaningful quota, refreshed weekly'. Not quantified.",
    "src": "https://antigravity.google/docs/plans/"
   },
   {
    "id": "cursor-plus",
    "vendor": "Cursor",
    "plan": "Plus",
    "price": 8,
    "kind": "credits",
    "credits": 8,
    "credits_hi": 80,
    "official": false,
    "caveat": "Cursor publishes the pool only as multiples of Plus. Est. low: worth at least its price (true of every $-credit plan we found). Est. high: ~10x, one user's measured ~$200 of API-priced tokens from a legacy $20 plan (Reddit).",
    "src": "https://www.reddit.com/r/cursor/comments/1udnfl2/using_composer_25_for_a_month_got_200_worth_of/"
   },
   {
    "id": "copilot-pro",
    "vendor": "GitHub Copilot",
    "plan": "Pro",
    "price": 10,
    "kind": "credits",
    "credits": 15,
    "official": true,
    "src": "https://github.com/features/copilot/plans"
   },
   {
    "id": "codex-go",
    "vendor": "OpenAI Codex",
    "plan": "Go",
    "price": 8,
    "kind": "window",
    "official": false,
    "models": [
     {
      "m": "GPT-6 Luna",
      "lo": 140,
      "hi": 3000
     }
    ],
    "caveat": "OpenAI publishes no Go numbers. Est.: low = Plus's GPT-6 Luna floor (350/5h) scaled by price (8/20); high = Plus's Luna ceiling.",
    "src": "https://developers.openai.com/codex/pricing/"
   },
   {
    "id": "jb-pro",
    "vendor": "JetBrains AI",
    "plan": "AI Pro",
    "price": 10,
    "kind": "credits",
    "credits": 10,
    "official": true,
    "caveat": "10 AI credits / 30 days, 1 credit = $1.",
    "src": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html"
   },
   {
    "id": "zed-pro",
    "vendor": "Zed",
    "plan": "Pro",
    "price": 10,
    "kind": "credits",
    "credits": 4.55,
    "official": true,
    "caveat": "$5 of hosted-model credit, billed at provider price +10% → $4.55 at API list price.",
    "src": "https://zed.dev/docs/account/plans-and-pricing"
   },
   {
    "id": "claude-pro",
    "vendor": "Claude",
    "plan": "Pro",
    "price": 20,
    "kind": "window",
    "official": false,
    "models": [
     {
      "m": "Claude Sonnet 5.5",
      "lo": 10,
      "hi": 40
     }
    ],
    "caveat": "Anthropic's 2025 estimate of 10–40 Claude Code prompts per 5h (archived help page). Anthropic no longer publishes counts and has raised 5-hour limits since (May and Sept 2026), so this is probably a floor.",
    "src": "https://archive.ph/G51jH"
   },
   {
    "id": "windsurf-pro",
    "vendor": "Windsurf",
    "plan": "Pro",
    "price": 20,
    "kind": "window",
    "unit": "day",
    "official": false,
    "models": [
     {
      "m": "Claude Opus 4.6",
      "lo": 7,
      "hi": 27
     }
    ],
    "caveat": "Windsurf's own Mar 2026 table: 7–27 'Premium Plus' (Opus 4.6-class) messages/day. Est.: one message = one Typical turn, slider = active days; weekly quota may bind.",
    "src": "https://devin.ai/blog/windsurf-pricing-plans"
   },
   {
    "id": "gemini-std",
    "vendor": "Gemini Code Assist",
    "plan": "Standard",
    "price": 22.8,
    "kind": "window",
    "unit": "day",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "Gemini 3.5 Flash",
      "lo": 1500,
      "hi": 1500
     }
    ],
    "caveat": "1,500 agent requests/day (Google). Est.: one request = one model call on the default Gemini 3.5 Flash; slider = active days.",
    "src": "https://developers.google.com/gemini-code-assist/resources/quotas"
   },
   {
    "id": "codex-plus",
    "vendor": "OpenAI Codex",
    "plan": "Plus",
    "price": 20,
    "kind": "window",
    "official": true,
    "models": [
     {
      "m": "GPT-6 Astra",
      "lo": 5,
      "hi": 45
     },
     {
      "m": "GPT-6.1 Sol",
      "lo": 15,
      "hi": 160
     },
     {
      "m": "GPT-6 Luna",
      "lo": 350,
      "hi": 3000
     }
    ],
    "caveat": "OpenAI's own local-message estimates per 5h. Weekly limits may also apply (not published).",
    "src": "https://developers.openai.com/codex/pricing/"
   },
   {
    "id": "kimi-plus",
    "vendor": "Kimi Code",
    "plan": "Plus",
    "price": 19,
    "kind": "window",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "Kimi K2.7 Code",
      "lo": 300,
      "hi": 1200
     }
    ],
    "caveat": "Kimi: a 5-hour window 'supports approximately 300–1,200 requests' (not broken down by plan); no weekly limit on new plans, but the shared monthly credit pool applies. Est.: requests = model calls.",
    "src": "https://www.kimi.ai/resources/kimi-code-introduction"
   },
   {
    "id": "zai-lite",
    "vendor": "Z.ai GLM",
    "plan": "Coding Lite",
    "price": 18,
    "kind": "credits",
    "credits": 86.96,
    "credits_hi": 173.91,
    "official": false,
    "caveat": "10,000 credits/week (Z.ai). Converted with Z.ai's own credit formula (GLM-5.3: 6.9 credits per 10K input tokens) at GLM-5.3's $1.40/M input → $0.00203/credit; high = all off-peak (0.5× credits). Weekly cap already applied.",
    "src": "https://docs.z.ai/devpack/overview"
   },
   {
    "id": "minimax-plus",
    "vendor": "MiniMax",
    "plan": "Token Plus",
    "price": 22,
    "kind": "window",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "MiniMax M2.7",
      "lo": 4500,
      "hi": 4500
     }
    ],
    "caveat": "4,500 M2.7 calls per 5h, captured from MiniMax's subscribe page (Sept 23) by a third party. Weekly window exists but isn't published.",
    "src": "https://iadecider.com/data/minimax-code-token-plan-2026-09.csv"
   },
   {
    "id": "ag-pro",
    "vendor": "Google Antigravity",
    "plan": "Google AI Pro",
    "price": 19.99,
    "kind": "window",
    "official": false,
    "models": [
     {
      "m": "Claude Opus 5.5",
      "lo": 8,
      "hi": 8
     }
    ],
    "cap_mo": 103,
    "caveat": "Google publishes no counts. Community test (Reddit): ~8 Opus prompts per 5h and ~24 a week, so we cap at ~103 turns/month. Low confidence.",
    "src": "https://www.reddit.com/r/google_antigravity/comments/1qmcqh0/antigravity_ai_pro_plan_opus_quota_test_results/"
   },
   {
    "id": "kiro-pro",
    "vendor": "Kiro",
    "plan": "Pro",
    "price": 20,
    "kind": "credits",
    "credits": 40,
    "official": false,
    "caveat": "1,000 credits × $0.04 add-on price. Credits are Kiro work units, not API dollars.",
    "src": "https://kiro.dev/docs/billing/"
   },
   {
    "id": "warp-build",
    "vendor": "Warp",
    "plan": "Build",
    "price": 20,
    "kind": "credits",
    "credits": 20,
    "official": true,
    "caveat": "'$20 of agent usage included each month at API rates'.",
    "src": "https://www.warp.dev/pricing"
   },
   {
    "id": "amp-ind",
    "vendor": "Amp",
    "plan": "Individual",
    "price": 20,
    "kind": "credits",
    "credits": 20,
    "official": false,
    "caveat": "Amp's page lists included agent usage without a $ figure; third-party trackers report ≥$20. Overage at provider prices, no markup.",
    "src": "https://www.softwr.com/pricing/amp-code"
   },
   {
    "id": "aug-std",
    "vendor": "Augment Code",
    "plan": "Standard",
    "price": 20,
    "kind": "credits",
    "credits": 14.29,
    "official": true,
    "caveat": "$20 included, but Augment adds a 40% service fee on LLM usage → $14.29 of API-priced tokens.",
    "src": "https://docs.augmentcode.com/models/token-based-pricing"
   },
   {
    "id": "factory-pro",
    "vendor": "Factory",
    "plan": "Pro",
    "price": 20,
    "kind": "credits",
    "credits": 14.7,
    "credits_hi": 56,
    "official": false,
    "caveat": "Factory now publishes only rolling limits. Est. from its earlier 20M Standard Tokens/month and Opus 5.5's 1.6× multiplier: low = cache reads count in full, high = cache reads at 10%.",
    "src": "https://deepwiki.com/factory-ai/factory/3.5-pricing-and-token-system"
   },
   {
    "id": "cursor-premium",
    "vendor": "Cursor",
    "plan": "Premium",
    "price": 30,
    "kind": "scale",
    "base": "cursor-plus",
    "factor": 4,
    "official": false,
    "caveat": "4x Plus (Cursor) × the Plus estimate.",
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   {
    "id": "copilot-proplus",
    "vendor": "GitHub Copilot",
    "plan": "Pro+",
    "price": 39,
    "kind": "credits",
    "credits": 70,
    "official": true,
    "src": "https://github.com/features/copilot/plans"
   },
   {
    "id": "windsurf-teams",
    "vendor": "Windsurf",
    "plan": "Teams (per seat)",
    "price": 40,
    "kind": "window",
    "unit": "day",
    "official": false,
    "models": [
     {
      "m": "Claude Opus 4.6",
      "lo": 7,
      "hi": 27
     }
    ],
    "caveat": "Same per-seat table as Pro (Windsurf, Mar 2026).",
    "src": "https://devin.ai/blog/windsurf-pricing-plans"
   },
   {
    "id": "jb-ult",
    "vendor": "JetBrains AI",
    "plan": "AI Ultimate",
    "price": 30,
    "kind": "credits",
    "credits": 35,
    "official": true,
    "caveat": "35 AI credits / 30 days (incl. $5 bonus), 1 credit = $1.",
    "src": "https://www.jetbrains.com/help/ai-assistant/licensing-and-subscriptions.html"
   },
   {
    "id": "kimi-pro",
    "vendor": "Kimi Code",
    "plan": "Pro",
    "price": 39,
    "kind": "window",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "Kimi K2.7 Code",
      "lo": 300,
      "hi": 1200
     }
    ],
    "caveat": "Same published 300–1,200 requests/5h range; Kimi doesn't publish a per-plan split, so higher tiers may get more.",
    "src": "https://www.kimi.ai/resources/kimi-code-introduction"
   },
   {
    "id": "kiro-proplus",
    "vendor": "Kiro",
    "plan": "Pro+",
    "price": 40,
    "kind": "credits",
    "credits": 80,
    "official": false,
    "caveat": "2,000 credits × $0.04.",
    "src": "https://kiro.dev/docs/billing/"
   },
   {
    "id": "ali-pro",
    "vendor": "Alibaba Model Studio",
    "plan": "Coding Plan Pro",
    "price": 50,
    "kind": "window",
    "unit": "month",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "Qwen3.7-Plus",
      "lo": 90000,
      "hi": 90000
     }
    ],
    "caveat": "90,000 requests/month (Alibaba; also 6,000/5h and 45,000/week). Est.: request = model call; cached price est. Limited stock, restocked daily.",
    "src": "https://www.alibabacloud.com/help/en/model-studio/coding-plan"
   },
   {
    "id": "minimax-max",
    "vendor": "MiniMax",
    "plan": "Token Max",
    "price": 55,
    "kind": "window",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "MiniMax M2.7",
      "lo": 15000,
      "hi": 15000
     }
    ],
    "caveat": "15,000 M2.7 calls per 5h (third-party capture of MiniMax's subscribe page). Weekly window unpublished.",
    "src": "https://iadecider.com/data/minimax-code-token-plan-2026-09.csv"
   },
   {
    "id": "zai-pro",
    "vendor": "Z.ai GLM",
    "plan": "Coding Pro",
    "price": 80,
    "kind": "credits",
    "credits": 521.74,
    "credits_hi": 1043.48,
    "official": false,
    "caveat": "60,000 credits/week, converted as for Lite.",
    "src": "https://docs.z.ai/devpack/overview"
   },
   {
    "id": "claude-max5",
    "vendor": "Claude",
    "plan": "Max 5x",
    "price": 100,
    "kind": "window",
    "official": false,
    "models": [
     {
      "m": "Claude Opus 5.5",
      "lo": 50,
      "hi": 200
     }
    ],
    "caveat": "Archived 2025 estimate (50–200 prompts/5h), scaled by Anthropic's current '5x Pro' multiplier. Probably a floor.",
    "src": "https://archive.ph/G51jH"
   },
   {
    "id": "codex-pro100",
    "vendor": "OpenAI Codex",
    "plan": "Pro 100",
    "price": 100,
    "kind": "scale",
    "base": "codex-plus",
    "factor": 5,
    "official": false,
    "caveat": "OpenAI's Codex lead on X: Pro 100 = 5x Plus (weekly; no 5-hour limit). × the Plus value.",
    "src": "https://www.seelig.ai/en/benchmarks/subscription-limits/chatgpt"
   },
   {
    "id": "cursor-super",
    "vendor": "Cursor",
    "plan": "Super",
    "price": 100,
    "kind": "scale",
    "base": "cursor-plus",
    "factor": 15,
    "official": false,
    "caveat": "15x Plus × the Plus estimate.",
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   {
    "id": "copilot-max",
    "vendor": "GitHub Copilot",
    "plan": "Max",
    "price": 100,
    "kind": "credits",
    "credits": 200,
    "official": true,
    "src": "https://github.com/features/copilot/plans"
   },
   {
    "id": "kimi-max",
    "vendor": "Kimi Code",
    "plan": "Max",
    "price": 99,
    "kind": "window",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "Kimi K2.7 Code",
      "lo": 300,
      "hi": 1200
     }
    ],
    "caveat": "Same published 300–1,200 range (no per-plan split).",
    "src": "https://www.kimi.ai/resources/kimi-code-introduction"
   },
   {
    "id": "ag-ultra5",
    "vendor": "Google Antigravity",
    "plan": "AI Ultra 5x",
    "price": 99.99,
    "kind": "scale",
    "base": "ag-pro",
    "factor": 5,
    "official": false,
    "caveat": "5x Pro allowance (Google) × the Pro community estimate.",
    "src": "https://antigravity.google/pricing/"
   },
   {
    "id": "kiro-promax",
    "vendor": "Kiro",
    "plan": "Pro Max",
    "price": 100,
    "kind": "credits",
    "credits": 200,
    "official": false,
    "caveat": "5,000 credits × $0.04.",
    "src": "https://kiro.dev/docs/billing/"
   },
   {
    "id": "aug-biz",
    "vendor": "Augment Code",
    "plan": "Business (team)",
    "price": 100,
    "kind": "credits",
    "credits": 71.43,
    "official": true,
    "caveat": "$100 included incl. 40% LLM fee → $71.43 of API-priced tokens.",
    "src": "https://docs.augmentcode.com/models/token-based-pricing"
   },
   {
    "id": "factory-plus",
    "vendor": "Factory",
    "plan": "Plus",
    "price": 100,
    "kind": "scale",
    "base": "factory-pro",
    "factor": 5,
    "official": false,
    "caveat": "~5x Pro (Factory) × the Pro estimate.",
    "src": "https://docs.factory.ai/pricing/individuals"
   },
   {
    "id": "claude-max20",
    "vendor": "Claude",
    "plan": "Max 20x",
    "price": 200,
    "kind": "window",
    "official": false,
    "models": [
     {
      "m": "Claude Opus 5.5",
      "lo": 200,
      "hi": 800
     }
    ],
    "caveat": "Archived 2025 estimate (200–800 prompts/5h). Price $200 (Claude Help Center). Probably a floor.",
    "src": "https://archive.ph/G51jH"
   },
   {
    "id": "codex-pro200",
    "vendor": "OpenAI Codex",
    "plan": "Pro 200",
    "price": 200,
    "kind": "scale",
    "base": "codex-plus",
    "factor": 10,
    "official": false,
    "caveat": "10x Plus for new subscriptions (existing: 20x until Oct 29), per OpenAI on X.",
    "src": "https://www.seelig.ai/en/benchmarks/subscription-limits/chatgpt"
   },
   {
    "id": "cursor-ultra",
    "vendor": "Cursor",
    "plan": "Ultra",
    "price": 200,
    "kind": "scale",
    "base": "cursor-plus",
    "factor": 50,
    "official": false,
    "caveat": "50x Plus × the Plus estimate.",
    "src": "https://cursor.com/docs/models-and-pricing"
   },
   {
    "id": "windsurf-max",
    "vendor": "Windsurf",
    "plan": "Max",
    "price": 200,
    "kind": "window",
    "unit": "day",
    "official": false,
    "models": [
     {
      "m": "Claude Opus 4.6",
      "lo": 42,
      "hi": 170
     }
    ],
    "caveat": "42–170 Premium Plus messages/day (Windsurf, Mar 2026).",
    "src": "https://devin.ai/blog/windsurf-pricing-plans"
   },
   {
    "id": "kimi-ultra",
    "vendor": "Kimi Code",
    "plan": "Ultra",
    "price": 199,
    "kind": "window",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "Kimi K2.7 Code",
      "lo": 300,
      "hi": 1200
     }
    ],
    "caveat": "Same published 300–1,200 range (no per-plan split).",
    "src": "https://www.kimi.ai/resources/kimi-code-introduction"
   },
   {
    "id": "zai-max",
    "vendor": "Z.ai GLM",
    "plan": "Coding Max",
    "price": 168,
    "kind": "credits",
    "credits": 1217.39,
    "credits_hi": 2434.78,
    "official": false,
    "caveat": "140,000 credits/week, converted as for Lite.",
    "src": "https://docs.z.ai/devpack/overview"
   },
   {
    "id": "minimax-ultra",
    "vendor": "MiniMax",
    "plan": "Token Ultra",
    "price": 132,
    "kind": "window",
    "calls": true,
    "official": false,
    "models": [
     {
      "m": "MiniMax M2.7",
      "lo": 15000,
      "hi": 36000
     }
    ],
    "caveat": "No published count. Est.: at least Max (15,000/5h), at most Max × price ratio (132/55).",
    "src": "https://platform.minimax.io/docs/guides/pricing-token-plan"
   },
   {
    "id": "ag-ultra20",
    "vendor": "Google Antigravity",
    "plan": "AI Ultra 20x",
    "price": 199.99,
    "kind": "scale",
    "base": "ag-pro",
    "factor": 20,
    "official": false,
    "caveat": "20x Pro allowance × the Pro community estimate.",
    "src": "https://antigravity.google/pricing/"
   },
   {
    "id": "kiro-power",
    "vendor": "Kiro",
    "plan": "Power",
    "price": 200,
    "kind": "credits",
    "credits": 400,
    "official": false,
    "caveat": "10,000 credits × $0.04.",
    "src": "https://kiro.dev/docs/billing/"
   },
   {
    "id": "warp-max",
    "vendor": "Warp",
    "plan": "Max",
    "price": 200,
    "kind": "credits",
    "credits": 200,
    "official": true,
    "caveat": "$200 of agent usage (10x Build).",
    "src": "https://www.warp.dev/pricing"
   },
   {
    "id": "factory-max",
    "vendor": "Factory",
    "plan": "Max",
    "price": 200,
    "kind": "scale",
    "base": "factory-pro",
    "factor": 10,
    "official": false,
    "caveat": "~10x Pro × the Pro estimate.",
    "src": "https://docs.factory.ai/pricing/individuals"
   }
  ]
 }
}