Anthropic's Most Capable Model
Anthropic's top model for demanding reasoning and long-horizon agentic work — a 1M-token context and adaptive thinking that holds focus across the hardest, multi-step builds. Put it to work in your browser, no setup required.
Model preset to Fable 5.1 · Sign up to unlock more models
The hottest practices from the X community right after launch: one-prompt games, 3D worlds, simulations, shaders and full apps people are building with Claude Fable 5.1.
Turned one short prompt into a fast, punchy Three.js shooter with real shooting feel, recoil and shader-driven graphics.
Apply to see the full prompt and try it
Make me the most insane and blast of a shooter you can possibly build on ThreeJS + Web shaders. The most important things are shooting, dynamics, recoil, and graphics.
From one prompt in 43 minutes, wrote a 1,660-line self-contained Frutiger Aero world in Three.js with a 600-frame camera fly-through.
Apply to see the full prompt and try it
Build a detailed Frutiger Aero world in a single self-contained HTML file using Three.js and nothing else, with a camera travelling through it for 600 frames. No external images.
Ran the same GTA-style open-world prompt through three models. Fable 5.1 nailed it in one go with no fixes: streets, traffic, pedestrians, a day-night cycle and drivable cars.
Apply to see the full prompt and try it
Build a GTA-style open world in one HTML file with zero assets and everything procedural: streets, traffic, pedestrians, a day-night cycle, and the ability to enter a car and drive.
Built an ARK-style survival game from a single prompt and played it multiplayer with friends, with cooking, crafting and 20+ hours of gameplay.
Build with Fable 5.1Ran a familiar game-build prompt through Fable 5.1 and got a strikingly complete result in a single shot.
Build with Fable 5.1Put Fable 5.1 head to head with another frontier model on the same prompt at max reasoning, comparing quality, time and cost side by side.
Build with Fable 5.1Generated a drowned neo-gothic tower as a twigl shader from the same prompt used on earlier models, showing a clear jump in visual quality.
Build with Fable 5.1Ran a RocketLeague-style game-design benchmark and got an impressive playable result in a two-shot prompt.
Build with Fable 5.1One-shot a playable medieval voxel kingdom with 2.4 million voxels, 1,500 soldiers and 160 working NPCs, including siege equipment and building interiors.
Build with Fable 5.1From one prompt, produced a raymarched ocean in hand-written GLSL with Gerstner waves, live foam, refracting glass, a volumetric sky and 30 sliders at 60fps.
Build with Fable 5.1Built a complete, playable game with Fable 5.1 from a single prompt.
Build with Fable 5.1Turned a single prompt into a fully functional engine in one shot, with zero follow-up iterations.
Build with Fable 5.1Ran the Boeing 747 benchmark and found Fable 5.1 clearly better than the previous version.
Build with Fable 5.1Built a neat little WW1 battlefield diorama from a single prompt, one of a series of tidy 3D scenes.
Build with Fable 5.1From one prompt, shipped a complete optical physics lab covering Bragg diffraction, beam waists and fringe visibility, with 285 passing browser tests and no hand-fixes.
Build with Fable 5.1Built the Battle of the Teutoburg Forest from a 500-word prompt: 15,724 soldiers, 4,626 trees and a 75-second battle, finishing its own work after running out of output budget.
Build with Fable 5.1Paired Fable 5.1 with a CLI agent and two custom skills to create a polished front-end website from one prompt.
Build with Fable 5.1One-shot a Cities: Skylines style city builder in Three.js on a clean install, spawning 14 agents to build its own version.
Build with Fable 5.1Had Fable 5.1 create a Black Ops 2 Zombies style remake from a single prompt, still going strong after 1.5 hours of work.
Build with Fable 5.1One-shot an entire Crossy Road style game from a quick, informal spoken prompt.
Build with Fable 5.1Generated a complete React app from a single prompt, covering layout, state and basic interactions in one pass.
Build with Fable 5.1One-shot an open-world game with NPCs, driving, combat, water physics and character animation, with no edits after generation.
Build with Fable 5.1Claude Fable 5.1 is Anthropic’s most capable model, released on September 1, 2026 as the successor to Claude Fable 5 (now a legacy model). Anthropic positions it as the top of the stack for demanding reasoning and long-horizon agentic work — the model to reach for when a task pushes past what other models can sustain. It has a 1M-token context window, up to 128K tokens of output, a June 2026 knowledge cutoff, and text & image input.
Anthropic recommends starting most workloads on Claude Opus 5 and reaching for Fable 5.1 only when Opus 5 at high effort falls short — Fable 5.1 trades a higher price for the extra headroom on the hardest reasoning, agentic-coding, and multistep research tasks. It keeps Fable 5’s $10 / $50 per-million-token pricing while cutting cache-read cost to a quarter, and adds stronger long-running agentic coding and document, spreadsheet, and slide work.
On Happycapy, Fable 5.1 isn’t just a chat box. It powers an agent inside a secure cloud sandbox that can write and run code, drive tools, and deliver finished work from a single prompt — with no local setup. Its adaptive thinking is always on, and an effort setting lets you dial depth against cost for each task.
Anthropic's published specs for the model, from platform.claude.com.
On Anthropic's own launch benchmarks, Fable 5.1 leads across agentic scientific research, agentic coding, computer use, reasoning and business workflows — ahead of Fable 5, Opus 5 and GPT-5.6 Sol. Switch tabs to see each eval.
Agentic scientific research. Fable 5.1 was evaluated with its production safeguards enabled.
Competitor figures are drawn from the respective developers' published system cards or benchmark leaderboards.
Source: Anthropic
| Fable 5.1 | Fable 5 | Opus 5 | GPT-5.6 Sol | |
|---|---|---|---|---|
| Agentic scientific research — Terminal-Bench-Science 0.1 | 52.6% | 24.7% | 29.0% | 22.4% |
| Agentic coding — Terminal-Bench 4.0 | 55.8% | 42.0% | 52.3% | 37.3% |
| Knowledge work — GDPval-AA v2 | 1853 | 1723 | 1824 | 1711 |
| Computer use — OSWorld 2.0 (partial) | 77.9% | 72.9% | 75.4% | — |
| Computer use — OSWorld 2.0 (strict) | 41.7% | 36.1% | 39.6% | — |
| Multidisciplinary reasoning — Humanity's Last Exam (no tools) | 60.9% | 57.8% | 56.6% | — |
| Multidisciplinary reasoning — Humanity's Last Exam (with tools) | 65.0% | 63.8% | 63.6% | — |
| Business workflows — AutomationBench | 31.4% | 17.1% | 26.9% | 19.6% |
| Agentic coding — CursorBench 3.2.0 | 73.4% | 70.5% | 70.0% | 67.2% |
Source: Anthropic. Best score per row in bold; a dash means no result was reported. Fable 5.1 was evaluated with its production safeguards enabled. On tasks where these safeguards intervened, Fable 5.1 and Fable 5 scored a zero on OSWorld 2.0, and Fable 5 scored a zero on AutomationBench. In all other safeguard interventions, cybersecurity tasks were completed by Claude Opus 4.8 and biology tasks by Claude Opus 5, which likely reduces the measured performance of Fable 5.1 and Fable 5 on these benchmarks.
Anthropic positions Fable 5.1 as its most capable model for the hardest reasoning tasks — the one to use when a lower-tier model at high effort still falls short.
Adaptive thinking and a 1M-token context let it sustain long, multi-step agent loops — planning, tool use, and iteration — without losing the thread.
Stronger long-running agentic coding than Fable 5: editing across files, running code, and driving toward a working result over extended runs.
Improved deep research and document, spreadsheet, and slide work — the data-heavy, professional tasks Anthropic highlights for the model.
Builders whose hardest tasks push past Opus 5 at high effort and need the extra headroom Fable 5.1 provides.
Anyone orchestrating long-running agents and workflows that benefit from a 1M-token context and always-on adaptive thinking.
Analysts and researchers running deep, multistep investigations and turning them into polished documents and decks.
People who want Anthropic’s frontier model for the toughest work and are willing to trade a higher price for the capability.
Claude Fable 5.1 is made by Anthropic. It was released on September 1, 2026 as the successor to Claude Fable 5, which is now a legacy model.
Yes. Anthropic positions Fable 5.1 at the top of its lineup for demanding reasoning and long-horizon agentic work. It still recommends starting most workloads on Claude Opus 5 and reaching for Fable 5.1 only when Opus 5 at high effort falls short.
Fable 5.1 keeps Fable 5’s $10 / $50 per-million-token pricing but cuts cache-read cost to a quarter, and improves long-running agentic coding, multistep research, and document, spreadsheet, and slide work. For developers calling the API directly there are a few breaking changes — forced tool use now errors, and thinking blocks are tied to the model that produced them — so review Anthropic’s migration notes before switching.
Fable 5.1 has a 1-million-token context window and can produce up to 128K tokens of output, large enough to work across entire codebases and long, multi-file projects.
Via the Claude API, Fable 5.1 is priced at $10 per million input tokens and $50 per million output tokens, with cache reads at $0.25 per million tokens. On Happycapy, Fable 5.1 is a premium model on paid plans; create a free account to explore the platform first, then upgrade to run it.
Yes. Fable 5.1 uses adaptive thinking that is always on, and exposes an effort setting (low, medium, high, xhigh, max) that defaults to high — so you can trade depth against speed and cost per task.
Use Claude Opus 5 for most work — Anthropic recommends it as the best-value flagship for agentic coding and enterprise tasks. Reach for Fable 5.1 when a task is hard enough that Opus 5 at high effort still falls short and the extra capability is worth the higher price.
Type your task into the prompt box on this page — the model is already set to Fable 5.1 — and hit send. Happycapy spins up a secure cloud sandbox and lets Fable 5.1 build, code, or analyze end-to-end, then delivers the result. No installs or API keys required.
Create a free account to explore Happycapy, then upgrade to put Anthropic's most capable model to work on your hardest reasoning, agentic, and research tasks — right in your browser, no setup needed.
Get started for free