Quick answer
Start most Claude workloads on Opus 5.5. Escalate to Fable 5.1 when higher Opus effort is not enough or a hard long-running task justifies the $10/$50 price.
- Released
- September 1, 2026
- API model
- claude-fable-5-1
- Price
- $10 in / $50 out per MTok; $0.25/MTok cache reads; 75% below Fable 5
- Context
- 1M tokens (128K max output)
01
Claude Fable 5.1: verified facts
| Field | Verified value |
|---|---|
| Vendor | Anthropic |
| Release | September 1, 2026 |
| API identifier | claude-fable-5-1 |
| Input / output price | $10 in / $50 out per MTok |
| Cache / fast pricing | $0.25/MTok cache reads; 75% below Fable 5 |
| Context / max output | 1M tokens (128K max output) |
| Availability | Current frontier option; defaults to high effort |
Every number on this page was checked against the linked vendor documentation on September 23, 2026. Prices are API list prices per million tokens; product subscriptions and regional availability are separate.
02
Where this model fits
Anthropic recommends Opus 5.5 for most workloads and Fable 5.1 as the higher-capability escalation. Fable keeps the $10/$50 base price of Fable 5 while cutting cache-read pricing to $0.25 per million tokens.
Your AI workspace for shipping software.
Download AIDEN free and point it at your existing Claude Code or Codex setup. No credit card, running in minutes.
Download AIDEN freeFree to start · macOS 12+ · No credit card required
03
When to use it
Use case 1
Demanding reasoning after evaluating Opus 5.5
Use case 2
Long autonomous coding and research runs
Use case 3
Large repositories that benefit from a 1M-token context window
| Strengths | Trade-offs |
|---|---|
| Anthropic's recommended step up when Opus 5.5 at higher effort falls short | Costs $10/$50 per MTok before caching |
| 1M-token context window and 128K maximum output | Slower than Opus 5.5 and Sonnet 5 |
| Cheaper cache reads than Fable 5 for long agentic sessions | Adaptive thinking is always on |
04
Published benchmarks
| Benchmark | Score | Source / caveat |
|---|---|---|
| Terminal-Bench 4.0 | 55.8% | Anthropic Opus 5.5 launch table, Vendor-published result |
| FrontierCode v1.1 | 50.3% | Anthropic Opus 5.5 launch table, Vendor-published result |
05