Last checked 2026-10-15 · 6 vendors · 24 plans · benchmarks as of Oct 2026

What your AI coding plan actually gets you.

Prices, usage windows, weekly caps and overages for Claude Code, Codex, Cursor, Copilot, Windsurf and Gemini. Each fact is sourced and dated, and every change is logged.

Compare by priceWhich plan fits me?

Same price, side by side

Plans are grouped by price so every row competes with plans that cost about the same. A green badge marks the best value in its column within that tier. Grey "not published" cells aren't counted. How we compare · All plans in detail

Free~$8–10~$20~$30–40~$100~$200

Free

Plan$/moIncluded usage$ credits per $1Best model includedTerminal-Bench 4.0AA Coding Agent IndexReal value ⓘ
Claude FreeChat only — Claude Code not includedsource ↗ Free 5-hour window, amount not published not published Sonnet (latest listed: 5.5)* 64.1% not published not published
GitHub Copilot Freesource ↗ Free 2,000 completions + 50 chat requests / mo not published Claude Opus 5.5 (listed) 65.2% ★ not published not published
Cursor Freesource ↗ Free Limited (not quantified) not published not stated (Agent + Chat, limited usage) not published not published not published
Windsurf Freesource ↗ Free Light daily+weekly quota (not quantified) not published not published not published not published not published

~$8–10

Plan$/moIncluded usage$ credits per $1Best model includedTerminal-Bench 4.0AA Coding Agent IndexReal value ⓘ
Cursor PlusOther providers' models: not available on Plussource ↗ $8 ★ Base weekly pool (1x) not published Grok 4.7 (pool also covers Grok 4.6, Grok 4.5, Composer 2.5) 28.8% 56.3Grok Build + Grok 4.7 (xhigh); Plus is fixed at medium effort not published
GitHub Copilot Prosource ↗ $10 $15 AI credits / mo $1.50$15 credits ÷ $10 Claude Opus 5.5 (listed) 65.2% ★ not published 1.5x+$15 credits/mo
OpenAI Codex Gosource ↗ $8 ★ Lightweight coding tasks; limits not published not published not published not published not published not published

~$20

Plan$/moIncluded usage$ credits per $1Best model includedTerminal-Bench 4.0AA Coding Agent IndexReal value ⓘ
Claude Prosource ↗ $20 ★ 5-hour window + weekly cap (≥5x Free per session; amounts not published) not published Opus (latest listed: 5.5)* 65.2% ★ 68.4 ★Claude Code + Sonnet 5.5 6.9x+$138–$551/mo API-equiv. · est.
Windsurf Prosource ↗ $20 ★ Daily+weekly quota (not quantified) not published Frontier OpenAI/Claude/Gemini (versions not stated) not published not published not published
Gemini Code Assist StandardNo new purchases since 2026-10-09source ↗ $22.8 1,500 agent/CLI requests / day (est. ~45,000 / 30-day mo) not published Gemini Pro/Flash (version not stated) not published not published not published
OpenAI Codex Plussource ↗ $20 ★ Per 5h (OpenAI estimate): 5–45 GPT-6 Astra, 15–160 GPT-6.1 Sol or 350–3,000 GPT-6 Luna local messages; weekly limits may apply not published GPT-6 Astra / GPT-6.1 Sol 59.6% 62.9Codex + GPT-6.1 Sol (xhigh) 19.7x+ ★$393–$3.5k/mo API-equiv. · official

~$30–40

Plan$/moIncluded usage$ credits per $1Best model includedTerminal-Bench 4.0AA Coding Agent IndexReal value ⓘ
Cursor PremiumOther providers' models: on-demand at API pricesource ↗ $30 ★ 4x Plus weekly pool; third-party models extra at API price not published Grok 4.7 (pool also covers Grok 4.6, Grok 4.5, Composer 2.5) 28.8% 56.3Grok Build + Grok 4.7 (xhigh) not published
GitHub Copilot Pro+source ↗ $39 $70 AI credits / mo $1.79$70 credits ÷ $39 Claude Opus 5.5 (listed) 65.2% ★ not published 1.8x+$70 credits/mo
Windsurf Teams (per seat)source ↗ $40 Pro-level daily+weekly quota not published not published not published not published not published

~$100

Plan$/moIncluded usage$ credits per $1Best model includedTerminal-Bench 4.0AA Coding Agent IndexReal value ⓘ
Claude Max 5xsource ↗ $100 5x Pro per 5-hour session + weekly cap (not published) not published Opus 5.5 / Fable* 65.2% ★ 68.4 ★Claude Code + Sonnet 5.5 13.8x+ ★$1.4k–$5.5k/mo API-equiv. · est.
OpenAI Codex Pro ($100)source ↗ $100 No 5-hour limit currently (OpenAI); weekly limits not quantified not published GPT-6 Astra 59.6% 62.9Codex + GPT-6.1 Sol (xhigh) not published
Cursor SuperOther providers' models: on-demand at API pricesource ↗ $100 15x Plus weekly pool not published Grok 4.7 (pool also covers Grok 4.6, Grok 4.5, Composer 2.5) 28.8% 56.3Grok Build + Grok 4.7 (xhigh) not published
GitHub Copilot Maxsource ↗ $100 $200 AI credits / mo (as shown on plans page) $2.00$200 credits ÷ $100 Claude Opus 5.5 (listed) 65.2% ★ not published 2.0x+$200 credits/mo

~$200

Plan$/moIncluded usage$ credits per $1Best model includedTerminal-Bench 4.0AA Coding Agent IndexReal value ⓘ
Claude Max 20xExact price not in fetched pagesource ↗ not published 20x Pro per 5-hour session + weekly cap (not published) not published Opus 5.5 / Fable* 65.2% ★ 68.4 ★Claude Code + Sonnet 5.5 $5.5k+$5.5k–$22.0k/mo API-equiv. · est.
OpenAI Codex Pro 20x ($200)source ↗ $200 No 5-hour limit currently (OpenAI); weekly limits not quantified not published GPT-6 Astra 59.6% 62.9Codex + GPT-6.1 Sol (xhigh) not published
Cursor UltraOther providers' models: on-demand at API pricesource ↗ $200 50x Plus weekly pool not published Grok 4.7 (pool also covers Grok 4.6, Grok 4.5, Composer 2.5) 28.8% 56.3Grok Build + Grok 4.7 (xhigh) not published
Windsurf Maxsource ↗ $200 Significantly higher quota (not quantified) not published not published not published not published not published

Which plan fits me?

This is a heuristic. No vendor publishes exact limits, so we rank plans that fit your budget and favor higher tiers for heavier use. Share the limits you hit to make it more accurate.

The weekly "what changed" email

Each week you get one email listing every price, cap or plan change, with sources.