Claude Fable 5.1
Claude Fable 5.1Rankingat high effort?
Full ranking →Alternatives?
CheaperGemini 4 Argonhighsame-49%*$1.99Synchronous output stops at 65,536+2BetterClaude Opus 5xhigh+6%+25%$4.88
OpenKimi K3max-4%-49%$2.00Sampling parameters are fixed+3
CheaperMuse Wonderxhigh-2%-60%$4 / $20Only auto tool_choice is supported+5
CheaperGPT-6 Astrahigh-2%-56%$1.73Monitoring can return HTTP 403+2Gemini 4 Argonhighsame-49%*$1.99Synchronous output stops at 65,536+2Muse Wonderxhigh-2%-60%$4 / $20Only auto tool_choice is supported+5GPT-6 Astrahigh-2%-56%$1.73Monitoring can return HTTP 403+2Kimi K3maxOpen-4%-49%$2.00Sampling parameters are fixed+3Claude Sonnet 5xhigh-4%-27%$2.87Claude Opus 5xhigh+6%+25%$4.88Kimi K3maxOpen-4%-49%$2.00Sampling parameters are fixed+3
GLM 5.3maxOpen-9%-49%$2.01No image input+4Muse Spark OpenxhighOpen-11%-78%$0.84Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1TInklingxhighOpen-38%-91%$0.95 / $4.05Context: 1M → 256K+1Claude Opus 5xhigh+6%+25%$4.88Claude Sonnet 5xhigh-4%-27%$2.87Claude Sonnet 5.5xhigh-23%-49%$2.01No forced tool use+5Claude Haiku 5.5high-43%-98%$0.079Remove sampling parameters+5Claude Haiku 4.5xhigh-72%—$1 / $5Context: 1M → 200KClaude Opus 5.5mediumn/a—$1.34No forced tool use+5Claude Mythos 5n/a—$10 / $50Claude Fable 5maxn/a—$8.75All Anthropic models →Benchmarks
in CodingWatch out 6
Replace forced tool_choice›
Forced tool_choice returns HTTP 400; use auto with an explicit instruction and strict tools, or JSON outputs. Source ↗
Adaptive thinking is always on›
Disabled thinking or manual budget_tokens returns HTTP 400; omit thinking or use adaptive thinking. Source ↗
No non-default sampling values›
Non-default temperature, top_p or top_k returns HTTP 400; omit non-default sampling values. Source ↗
Requires 30-day data retention · Claude API›
A Claude API organization or workspace without 30-day retention receives HTTP 400; configure workspace retention or contact the Anthropic account team. Source ↗
Thinking blocks bind to conversations›
Where conversation binding is enforced, replaying an invalidated block returns HTTP 400; use the documented drop_block control to continue without it. Source ↗
No assistant prefill›
Assistant prefill returns HTTP 400; use system prompt instructions instead. Source ↗
FAQ
How much does Claude Fable 5.1 cost?›
$10 per 1M input tokens and $50 per 1M output tokens; cached input $0.25 per 1M. A typical benchmark task costs about $3.91 at high effort.
Is Claude Fable 5.1 free to use?›
There is no free API; the cheapest paid price is $10 per 1M input tokens.
How good is Claude Fable 5.1, and what is it best at?›
#3 of 334 in Chat & Knowledge, #7 of 132 in Coding, #30 of 133 in Agents & Tool Use, #7 of 195 in Writing & Language, #5 of 37 in Vision & Documents. Strongest at: Coding, Agents & work tasks, Reasoning, Science, Long context, Knowledge.
Is Claude Fable 5.1 better than Claude Opus 5?›
On SWE-bench Pro (v2 hard), Claude Fable 5.1 scores 92.2 vs 98 — lower. It is 20% cheaper per task.
How can I use Claude Fable 5.1?›
Through the API of Amazon Bedrock, Anthropic, Azure, Google.
What is the context window of Claude Fable 5.1?›
1M tokens, with up to 128K output tokens.
When was Claude Fable 5.1 released? Is there a newer version?›
Released 2026-09-01. It is the current version in its line.
Is Claude Fable 5.1 open source?›
No, Claude Fable 5.1 is a closed model available through an API. The closest open-weights option here is Kimi K3.
Can Claude Fable 5.1 read images?›
It accepts text and images, PDF files as input.
What is Claude Fable 5.1?›
Claude Fable 5.1 is a closed language model from Anthropic, released Sep 2026, the successor to Claude Fable 5.
What are the best alternatives to Claude Fable 5.1?›
Gemini 4 Argon (cheaper), Claude Opus 5 (better), Kimi K3 (open weights).