Updated on November 10, 2026
Claude Fable 5.2 is Anthropic's most capable large language model available to the general public, released on November 9, 2026 alongside its restricted counterpart Claude Mythos 5.2. API pricing sits at $8 per million input tokens and $40 per million output tokens, a 20% reduction from Fable 5.1. The model targets coding, finance, scientific research and long-horizon agent orchestration. No free plan grants access; a Pro subscription ($20/mo), Max, Ultra, Team, Enterprise or an API key is required.
Fable 5.2 and Mythos 5.2 share the same underlying model. Fable adds cyber and biology safeguards, making it available to every paying subscriber, while Mythos stays locked behind Anthropic's verification programs. According to Anthropic, Fable 5.2 blocks roughly 55% fewer queries than Fable 5.1 (about 60% fewer in cyber, 40% in biology). Blocked requests fall back to Opus 4.8 or Opus 5 automatically.
Benchmarks published by Anthropic paint a clear picture: 49.1% on FrontierCode 2.0 at its highest effort setting, ahead of Opus 5.5 (45.5%), GPT-6 Astra (42.8%) and Gemini 4 Carbon (41.1%). On Terminal-Bench-Science the model reaches 74.6%, and a swarm of 1,000 Fable 5.2 agents formalized the proof of Fermat's Last Theorem in Lean in 30 hours, down from 11 days for the first attempt.
Fable 5.2 does not dominate everywhere. Gemini 4 Carbon leads on AutomationBench (53.9% vs 51.4%), and Opus 5.5 keeps a higher Elo on the AA-Briefcase office-work test (1,800 vs 1,785) at half the token price. Anthropic itself recommends starting with Opus 5.5 for everyday work and stepping up to Fable only when the problem demands it.
The prices below are the dollar figures published by Anthropic at launch. They may change; always check the official pricing page.
| Mode | Input ($/M tokens) | Output ($/M tokens) |
|---|---|---|
| Standard | 8 | 40 |
| Fast mode | 16 | 80 |
| Batch API (−50%) | 4 | 20 |
| Cache (read) | 0.20 | — |
| Opus 5.5 (comparison) | 4 | 20 |
Anthropic filed its S-1 with the SEC on October 26, 2026, then amended it on November 2 to include third-quarter revenue of $16.96 billion, according to the filing. The planned ticker is ANT on the Nasdaq Global Select Market. Valuation expectations, initially floated at roughly $2 trillion, were walked down by investors: Bloomberg reported on November 6 a range of $1.2 to $1.3 trillion, with some buyers as low as $1.1 trillion.
Fable 5.2's release falls squarely in that window, as Bloomberg reported that Anthropic could begin its IPO roadshow as early as the week of November 9, a timeline the company has not confirmed. Among competing LLM models, GPT-6 Astra and Gemini 4 Carbon remain the closest rivals; each outperforms Fable 5.2 on at least one benchmark, keeping the choice firmly task-dependent.
Claude Fable 5.2 is Anthropic's most capable publicly available model, released November 9, 2026. It shares the same base as Mythos 5.2 but includes cyber and biology safeguards. With a one-million-token context window and 128K-token max output, it targets advanced coding, finance, scientific research and multi-hour agent orchestration.
Standard API pricing is $8 per million input tokens and $40 per million output tokens, 20% below Fable 5.1. Fast mode doubles that to $16/$80, and the batch API halves it to $4/$20. On claude.ai, Max ($100/mo) and Ultra ($200/mo) include Fable up to half of weekly limits, while Pro uses usage credits.
No free access exists. The Free plan excludes Fable entirely. Max ($100/mo) and Ultra ($200/mo) include it up to half the weekly usage limit. Pro subscribers can use it through usage credits. At launch, eligible paid subscribers received one to three extra Fable quota resets depending on account history.
Anthropic recommends Opus 5.5 ($4/$20) for most tasks and Fable 5.2 for the hardest problems. Fable leads on FrontierCode 2.0 (49.1% vs 45.5%) and Remote Labor Index (41.3% vs 21.3%), but Opus 5.5 edges ahead on AA-Briefcase Elo (1,800 vs 1,785) at half the per-token cost. Start with Opus, escalate to Fable when the difficulty warrants double the spend.
Verdict: At $8 per million input tokens and top-of-chart coding benchmarks according to Anthropic, Fable 5.2 earns its place on the shortlist of developers, R&D teams and analysts tackling problems that Opus 5.5 cannot crack alone.
« An all-in-one creative platform that brings together more than 30 image and video models in a single interface. It includes camera controls that let you direct like a real filmmaker, characters that remain consistent from scene to scene, and ready-to-use tools for ads and short-form content »
Claude Opus 5.5 
Gemini 4 Carbon 
GPT-6 Astra 
