Need a China coding plan (GLM / MiniMax / Kimi / Volcengine / Xiaomi)? Email me at [email protected]
Anthropic · Claude Code
The gold standard for AI coding. Pro at $20/mo gives you Claude Opus 5.5, Sonnet 5 and Haiku 4.5 with extended thinking in the terminal, IDE and web; Max 5x and 20x multiply usage 5x-20x and include Fable 5.1 at 50% of weekly limits.
Current lineup: Fable 5.1 and Opus 5.5 (both shipped in September 2026), plus Sonnet 5 and Haiku 4.5. Opus, Sonnet and Haiku are included on every paid tier; Fable 5.1 needs usage credits on Pro and is capped at 50% of weekly limits on Max.
New in September 2026 (API name claude-fable-5-1): Anthropic’s top-tier model for long-running agents at $10/$50 per million tokens, with cache reads down 75% to $0.25 per million and High effort as the default in Claude Code. Not included on Free; needs usage credits on Pro and 50% of weekly limits on Max.
New flagship for complex agentic coding and enterprise work (API name claude-opus-5-5): $4 input / $20 output per million tokens, 20% below Opus 5, with cache reads at $0.20 per million and 30%+ faster output generation. Fast mode ($8/$40) runs up to 2.5x faster in Claude Code.
High-performance coding and agent model (released 2026-06-30). Balanced speed and capability for daily workloads.
Fastest and cheapest tier, ideal for quick iterations and code completion.
Pro for individuals, Max for heavy usage with 5x-20x multipliers and priority access. Usage runs on a 5-hour rolling window plus a weekly cap on paid tiers. Annual billing saves ~15%; Free covers Sonnet 5 and Haiku 4.5 only, with no Fable 5.1 access.
Pro
$17/mo billed annually (-15%)
Individual developers who want the best model quality for daily coding
Max 5x
Popular5x Pro usage
Heavy users who hit Pro limits regularly
Max 20x
Top tier20x Pro usage
Professional teams and extremely heavy usage patterns
Usage is measured relative to Pro baseline limits on a 5-hour rolling window plus a weekly cap.
| Tier | Price | Usage | Peak priority | Fable 5.1 | Early features |
|---|---|---|---|---|---|
| Pro | $20/mo | 1x | No | Usage credits | No |
| Max 5x | $100/mo | 5x | Yes | 50% of weekly limits | No |
| Max 20x | $200/mo | 20x | Yes | 50% of weekly limits | Yes |
Opus 5.5 and Fable 5.1 rank at the top of coding benchmarks for agentic tasks.
Model Context Protocol lets you plug in tools, memory and data sources.
Opus 5.5 reasons step by step on complex problems before answering; Fable 5.1 defaults to High effort in Claude Code.
Fable 5.1 and Opus 5.5 shipped in September and are available across platforms. Claude Mythos 5.1 stays limited to vetted security and life-science teams at US institutions only.
Pro ($20/mo) covers individual usage with Opus 5.5, Sonnet 5 and Haiku 4.5. Max 5x ($100/mo) gives 5x Pro usage with priority access; Max 20x ($200/mo) gives 20x usage, early feature access and peak-hour priority. Fable 5.1 needs usage credits on Pro and runs at 50% of weekly limits on both Max tiers. Annual billing saves about 15% on Pro.
Pro and Max include Claude Opus 5.5, Sonnet 5 and Haiku 4.5; Free covers Sonnet and Haiku only. Fable 5.1 (top tier for long-running agents, $10/$50 per million tokens) is not part of any plan: on Pro you pay for it with usage credits, and on Max 5x / Max 20x it draws on 50% of your weekly limits.
Yes. Claude Code runs in the terminal, VS Code, JetBrains IDEs, and as Claude Cowork on the web and desktop. Claude in Chrome is also generally available for paid plans.
Yes, Free costs $0 and covers unlimited chat with basic models, code/data visualization, web search and cross-conversation memory — but model access is limited to Sonnet and Haiku, with no Opus 5.5 or Fable 5.1. The Pro plan at $20/mo is the standard entry point for Claude Code.
Anthropic generally does not refund activated subscriptions. Start with Pro to evaluate before upgrading to Max.