Claude Fable 5.2: What's New, What It Costs and Who Needs It

Anthropic released Claude Fable 5.2 on Monday, 9 November 2026, alongside a sister model called Claude Mythos 5.2. It is the company's most capable model that anyone can use, and it arrives with a price cut, a faster output speed and a long list of benchmark claims. Here is what changed, what the numbers actually say, and who should care.
What Fable 5.2 is
Fable and Mythos share the same underlying model. The difference is the safeguards. Fable is the public version, with protections around cybersecurity and biology. Mythos has fewer restrictions and is available only to vetted organisations through Anthropic's Life Sciences and Cyber Verification Programs, and through Claude Security on Enterprise plans. Fable 5.2 is available on all paid Claude plans, on the Claude Platform and on the major clouds, including AWS, Google Cloud and Microsoft Azure.
Anthropic's own pitch is that the model suits hard, open-ended technical work. A company spokesperson told CNBC it is meant for power users and power use cases such as engineering, mathematics and robotics, while Opus 5.5 remains the better fit for everyday tasks.
Speed, price and defaults
- Speed: Anthropic says outputs are generated up to 30 percent faster than with the 5.1 versions. A separate fast mode, available in Claude Code and on some platforms, promises up to 2.5 times faster token generation.
- Price: the API price falls from $10 per million input tokens and $50 per million output tokens to $8 and $40. That is a 20 percent cut, and Fable still costs more than Opus.
- Efficiency: at medium effort the model uses fewer tokens than Fable 5.1 on hard tasks. Medium is now the default effort level in the Claude apps and Claude Code.
- Fewer blocks: Anthropic reports roughly 55 percent fewer blocked requests than Fable 5.1 in early testing, while saying the cyber and biology safeguards are as robust as before.
The benchmark claims
The headline number is on FrontierCode 2.0, a coding benchmark run by Anthropic. Fable 5.2 scored 49.1 percent at its highest effort setting. For comparison, Anthropic reports 32.4 percent for Fable 5.1, 45.5 percent for Opus 5.5 and 42.8 percent for GPT-6 Astra. At high effort the company says Fable 5.2 matches Opus 5.5 for about half the cost per task.
The company also says the model more than doubled the scores of GPT-6 Astra and Fable 5.1 on an internal robotics simulation test, and that it solved two open problems in black hole physics during testing. These are results published by the company that built the model, and most were run internally. Independent testing usually lands somewhere different, so treat them as a starting point rather than a verdict.
Why the launch drew attention
Two threads explain the interest. The first is timing: CNBC reported that these are the first frontier models Anthropic has released since its chief executive called for slowing the pace of AI development, and that the release comes as the company prepares for a possible IPO. An Anthropic executive described the approach as pacing rather than pausing, according to Bloomberg.
The second is history. Fable 5 first launched on 9 June 2026. Three days later, on 12 June, the US Commerce Department ordered Anthropic to suspend access to Fable 5 and Mythos 5 for any foreign national, which in practice took the models offline worldwide. The controls were lifted on 30 June, and access began returning on 1 July with stronger safeguards. Fable 5.1 followed on 1 September. Against that background, a smooth 5.2 release matters more than a typical version bump.
Who should use it
For most people the answer is modest. If you mainly write emails, summarise documents or answer questions, a cheaper model will do the job. Fable 5.2 makes sense for long coding sessions, hard maths, multi-step agent work and research where a better answer is worth the extra cost. Subscribers on eligible paid plans get extra Fable usage this week, and Fable counts toward up to half of weekly limits on Pro plans.
One naming note: Anthropic's Opus, Sonnet and Haiku models are now at version 5.5, while Fable and Mythos sit at 5.2. The New Stack called that a mild surprise. Version numbers across different model families do not map onto each other, so the gap says little about quality.
Sources: reporting by CNBC, MacRumors, Android Authority, The New Stack, The Next Web, Yahoo Finance and ITPro, plus Anthropic's announcement. Benchmark figures are reported by Anthropic and were not independently verified.
تعليقات
إرسال تعليق