AI news6 min read
Claude Fable 5.2 and Mythos 5.2 arrive at a lower price
Anthropic’s top models get faster and cheaper, at $8 and $40 per million tokens, and lead most of its own benchmark table in coding, science and computer use.
In this story7 parts
Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, upgrades to the models it reserves for its hardest technical work. Anthropic says they generate output up to 30% faster than the 5.1 versions, and they cost less: $8 per million input tokens and $40 per million output tokens.
Fable and Mythos are the same underlying model. Fable has strong cybersecurity and biology safeguards so it can be offered to everyone. Mythos, with fewer of those restrictions, is available only to vetted organizations through Anthropic’s Cyber and Life Sciences Verification Programs.
Anthropic’s launch video
What Anthropic says changed
- Faster and cheaper. Output arrives up to 30% faster than Fable 5.1 and Mythos 5.1, at a lower price of $8 and $40 per million input and output tokens. Our Fable 5.1 coverage listed $10 and $50.
- Fast mode. An optional mode speeds up token generation by up to 2.5 times, at $16 and $80 per million tokens.
- Fewer tokens at the default setting. At Medium effort, the default in the Claude apps and Claude Code, Anthropic says Fable 5.2 completes hard tasks with fewer tokens than 5.1.
- Clearer communication. Anthropic says testers found the new models explain themselves more clearly.
- Safeguards with fewer false blocks. Fable 5.2’s cyber and biology safeguards are, Anthropic says, as robust as 5.1’s but block fewer legitimate requests.
Anthropic’s guidance on when to use it is worth noting: for most coding and knowledge work, it recommends starting with Claude Opus 5.5 and moving up to Fable or Mythos only for the most ambitious, open-ended work. Fable 5.2 can also act as an “advisor” that smaller Claude models consult, or as an “orchestrator” that hands simpler jobs to them.
The benchmarks
From Anthropic’s launch table. These are Anthropic’s own figures; scores use each model’s highest available effort unless noted, with safeguards on for the Claude models.
| Benchmark | Fable 5.2 | Fable 5.1 | Opus 5.5 | GPT-6 Astra | Gemini 4 Carbon |
|---|---|---|---|---|---|
| FrontierCode 2.0 (agentic coding) | 49.1% | 32.4% | 45.5% | 42.8% | 41.1% |
| FrontierSWE v2 (agentic coding) | 69.9% | 56.3% | 62.3% | 65.5% | 61.2% |
| AutomationBench (business workflows) | 51.4% | 31.4% | 42.5% | 41.4% | 53.9% |
| Terminal-Bench-Science 0.1 | 74.6% | 52.6% | 58.7% | 68.1% | 64.8% |
| OSWorld 2.1 (computer use) | 89.3% | 83.5% | 87.2% | 79.5% | 81.5% |
| WANDR (agentic search) | 82.1% | 68.7% | 72.3% | — | — |
| RoboDojo Sim (robotics) | 64.9% | 27.6% | 39.4% | 29.8% | — |
| Remote Labor Index (knowledge work) | 41.3% | 17.9% | 21.3% | 20.8% | — |
| Humanity’s Last Exam (with tools) | 74.2% | 65.6% | 67.7% | 57.2% | — |
| Chartography (no tools) | 73.4% | 44.8% | 64.4% | 71.0% | 72.5% |
| AA-Briefcase v1.1 (knowledge work, Elo) | 1,785 | 1,669 | 1,800 | 1,569 | 1,611 |
Three points stand out:
- It is not first everywhere. Gemini 4 Carbon leads on AutomationBench (53.9% to 51.4%), and Opus 5.5 is slightly ahead on the AA-Briefcase Elo score.
- The biggest jumps are in long, open-ended work. Fable 5.2 more than doubles Fable 5.1 on robotics simulation and the Remote Labor Index.
- Cost efficiency. Anthropic says Fable 5.2 at High effort roughly matches Opus 5.5’s best FrontierCode score (45.3% vs. 45.5%) for about half the cost per task, and at Medium effort it matches Opus 5.5’s top WANDR score for about 40% of the cost.
Anthropic’s table also reports Mythos 5.2 on specialist tests: 76.3% on CyScenarioBench and 89.6% on its internal CVE Rediscovery evaluation (cybersecurity), and 85.1% on a 24-hour protein binder design task. Anthropic notes that on CyScenarioBench each lab’s scores come from different tasks, so they are not directly comparable.
What it did in testing
Anthropic says that during pre-release testing the models helped its engineers complete software migrations that were otherwise infeasible, designed three working electronic devices from scratch, solved two open problems in the physics of black holes, and formalized the proof of Fermat’s Last Theorem in 30 hours. In one experiment, Fable 5.2 coordinated more than 1,000 subagents over eight days to write a prototype operating system that boots on a laptop.
Here is one of Anthropic’s demos, in which Fable 5.2 designs a working voice-assistant hub:
These are Anthropic’s showcase examples. Independent testing has not caught up yet.
Availability and price
- API:
claude-fable-5-2on the Claude Platform, plus Amazon Web Services, Google Cloud and Microsoft Azure. $8 and $40 per million input and output tokens; fast mode $16 and $80. - Claude plans: This week Anthropic is giving eligible paid subscribers extra Fable usage: one to three Fable limit resets depending on how long they have subscribed. Fable is also included for up to 50% of weekly usage limits on Pro plans and standard Team seats. Our guide to Claude and ChatGPT usage limits explains how Fable counts toward plan limits.
- Mythos 5.2: for cybersecurity professionals through the Cyber Verification Program, and for life-sciences researchers through the Life Sciences Verification Program. Claude Enterprise customers can also use it through Claude Security.
- Data retention: Anthropic says these models rely on safety monitoring across a window of usage. Standard customers get 30-day retention for that monitoring. A new option, Enterprise Frontier Safeguards, runs the monitoring inside a customer’s own infrastructure so it works with zero data retention. It is in private beta.
What it means for working professionals
- Engineers: Fable 5.2 is aimed at long, multi-step jobs such as migrations, audits and large refactors. For everyday coding, Anthropic itself points to Opus 5.5 first.
- Analysts and knowledge workers: the gains on the Remote Labor Index and chart reading point to better long-form office work, but Opus 5.5 still edges it on one knowledge-work measure.
- Budget holders: the per-token price fell 20%, but long agent runs still add up. Watch cost per finished task rather than the headline rate.
- Security and life-sciences teams: the strongest capabilities sit in Mythos 5.2, which requires verification.
Everything above comes from Anthropic’s announcement and videos. We have not tested Fable 5.2 ourselves.
Sources
Every story names its source. Check our work.