# Claude Fable 5.1 Claude Fable 5.1 is an AI model released by Anthropic on Sep 1 2026. It has a 1M context window. At release it scored 65% on Humanity's Last Exam (with tools), 81.2% on SWE-Bench Pro and 52.6% on Terminal-Bench-Science 0.1. ## Facts | Field | Value | | --- | --- | | Model | Claude Fable 5.1 | | Developer | Anthropic | | Release date | Tuesday, Sep 1 2026 | | Licensing | Proprietary | | Context window | 1M | ## API pricing All rates in USD per 1,000,000 tokens, pay-as-you-go. | Tier | Input | Cached input | 5 min cache write | 1 hr cache write | Output | | --- | --- | --- | --- | --- | --- | | All documented contexts | $10.00 | $0.25 | $12.50 | $20.00 | $50.00 | Cache reads are priced at 0.025x the base input price on this model, against 0.1x on every other Claude model. Verified September 1, 2026 against the first-party source: https://platform.claude.com/docs/en/about-claude/pricing ## Benchmark scores published at release | Benchmark | Score | Source | What it measures | | --- | --- | --- | --- | | SWE-Bench Pro | 81.2% | Lab | Can the AI fix real bugs in real software? It's handed actual problems from open-source projects and has to write code that genuinely solves them. Higher is better. | | SWE-Bench Multilingual | 89.1% | Lab | Like SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | | SWE-Bench Multimodal | 54.7% | Lab | Real bug reports that arrive with pictures attached — a screenshot, a mockup, a page rendering wrongly — so the AI has to read the image as well as the code to work out what to fix. Higher is better. | | Terminal-Bench 4.0 | 55.8% | Lab | Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Version 4.0 recalibrated how much time, CPU and memory each task gets, removed eight tasks and fixed nineteen, so fewer runs fail for reasons that have nothing to do with the model. Scores are not comparable with earlier versions. Higher is better. | | Terminal-Bench-Science 0.1 | 52.6% | Lab | The same command-line setup as Terminal-Bench, pointed at scientific work: the AI has to drive research tooling and computational workflows through to a result, rather than administer a machine. Version 0.1 is the first release of the task set, and scores run lower than on the general board. Higher is better. | | Humanity's Last Exam (no tools) | 60.9% | Lab | Humanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better. | | Humanity's Last Exam (with tools) | 65% | Lab | Humanity's Last Exam — extremely hard expert questions across many subjects. “With tools” means the AI is allowed to search the web or run code while answering. Higher is better. | | ARC-AGI-2 | 90% | Lab | Puzzle-style tests of abstract reasoning and pattern-finding — the kind of thing people find easy but AIs often struggle with. Higher is better. | | AutomationBench | 31.4% | Lab | Tests whether the AI can run real multi-step business workflows — the kind of end-to-end office processes companies want to automate — from start to finish. Higher is better. | | HealthBench Professional | 62.1% | Lab | Realistic health conversations graded against detailed rubrics written by physicians — can the AI respond the way a careful medical professional would? Higher is better. | | GDPval-AA v2 | 1853 | Lab | economically valuable knowledge work (v2, re-based Elo) | | AA-Briefcase | 1694 | Lab | Artificial Analysis agentic office-work eval (Elo) | ## About Claude Fable 5.1 Claude Fable 5.1, released September 1, 2026, succeeded Claude Fable 5 at the top of Anthropic's lineup twelve weeks after the Claude 5 generation opened. It shipped at the same $10 per million input tokens and $50 per million output as its predecessor, with one price cut: cache reads dropped to $0.25 per million, a quarter of Fable 5's rate and 0.025x the base input price where every other Claude model charged 0.1x. Anthropic described it as its most capable widely released model and pitched it narrowly — start with Claude Opus 5, and move up to Fable 5.1 for demanding reasoning and long-horizon agentic work, or when Opus 5 at its highest effort still falls short. The launch notes concentrated the gains in six areas: agentic coding over sessions that run for hours, knowledge work that carries an analysis through to a finished document, spreadsheet or slide deck, multistep research, reading dense charts and tables inside PDFs, reasoning across the full 1M-token context window, and computer use. Adaptive thinking was always on, with a 128K output limit and a June 2026 knowledge cutoff. It also brought breaking API changes: forced tool use returned an error, and thinking blocks became bound to the model and the conversation that produced them. Every response carried Anthropic's text watermark, and media files came with C2PA Content Credentials. Claude Mythos 5.1 shipped the same day as the Project Glasswing twin. The launch table, run with production safeguards enabled, put Fable 5.1 at 55.8% on Terminal-Bench 4.0 against Fable 5's 42.0% and Opus 5's 52.3% under the same harness, 52.6% on Terminal-Bench-Science 0.1 (more than double Fable 5's 24.7%), a GDPval-AA v2 Elo of 1853, 60.9% on Humanity's Last Exam without tools and 65.0% with, and 31.4% on AutomationBench. Its OSWorld 2.0 figures, 77.9% partial and 41.7% strict, were run on the benchmark's August 2026 task release, which Anthropic said made them not directly comparable to earlier published scores. The system card filled in the coding detail the launch post left out: 81.2% on SWE-bench Pro, 89.1% on SWE-bench Multilingual and 54.7% on SWE-bench Multimodal, alongside an AA-Briefcase Elo of 1694 and 62.1% on HealthBench Professional. ## Questions and answers ### When was Claude Fable 5.1 released? Claude Fable 5.1 was released by Anthropic on Tuesday, Sep 1 2026. ### Who made Claude Fable 5.1? Claude Fable 5.1 was built by Anthropic. AI safety company building the Claude family of models. Founded in 2021 by former OpenAI researchers. ### How much does Claude Fable 5.1 cost? Claude Fable 5.1 costs $10.00 per million input tokens and $50.00 per million output tokens through the Anthropic API. Cached input is $0.25 per million tokens. Rates are pay-as-you-go API prices verified against Anthropic's published pricing on September 1, 2026. ### What benchmark scores did Claude Fable 5.1 get? Claude Fable 5.1 reports 12 tracked benchmark scores — SWE-Bench Pro: 81.2%; SWE-Bench Multilingual: 89.1%; SWE-Bench Multimodal: 54.7%; Terminal-Bench 4.0: 55.8%; Terminal-Bench-Science 0.1: 52.6%; Humanity's Last Exam (no tools): 60.9%; Humanity's Last Exam (with tools): 65%; ARC-AGI-2: 90%; AutomationBench: 31.4%; HealthBench Professional: 62.1%; GDPval-AA v2: 1853; AA-Briefcase: 1694. Scores are the figures published at release by Anthropic. It holds the best score among all models tracked here on SWE-Bench Pro, Terminal-Bench-Science 0.1, Humanity's Last Exam (no tools), Humanity's Last Exam (with tools) and AA-Briefcase. ### What is the context window of Claude Fable 5.1? Claude Fable 5.1 has a context window of 1M. That is the maximum amount of input plus output the model can hold in a single request. ### Is Claude Fable 5.1 open source? No. Claude Fable 5.1 is a proprietary model. The weights are not published — it is available only through the provider's own API, apps, or partner platforms. ### What came before and after Claude Fable 5.1? Anthropic's previous tracked release was Claude Opus 5 on Jul 24 2026, 39 days earlier. It was followed by Claude Mythos 5.1 on Sep 1 2026. --- Canonical page: https://aireleasetracker.com/model/anthropic/claude-fable-5.1 Full dataset: https://aireleasetracker.com/llms-full.txt · JSON: https://aireleasetracker.com/models.json Source: AI Release Tracker (https://aireleasetracker.com). Benchmark scores are the figures published by the releasing lab at launch.