Anthropic Launches Claude Fable 5.2 and Mythos 5.2, Cutting API Prices 20% and Speeding Outputs 30%

• Dispatched: Nov 9, 2026 · 19:38 ET (Nov 10, 00:38 UTC)

One-sentence summary: Anthropic launched Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026 — two versions of the same underlying model, with Fable open to all users and Mythos restricted to vetted organizations — cutting standard API prices to $8 per million input tokens and $40 per million output tokens, speeding outputs by up to 30% versus version 5.1, and beating Opus 5.5 on most published benchmarks.

Verification status: ✓ Primary disclosure from Anthropic's official announcement, corroborated by multiple outlets.

Core Facts & Consensus Delta:

Anthropic announced the models at 2026-11-09 14:38 ET / 19:38 UTC. Fable 5.2 carries cybersecurity and biology safeguards and is available to all users; Mythos 5.2 lifts some of those safeguards and is available only to organizations admitted to Anthropic's Cyber Verification and Life Sciences Verification Programs, plus enterprise customers through Claude Security.

Pricing falls 20% from version 5.1: standard API rates are $8 per million input tokens and $40 per million output tokens, versus $10 and $50 for Fable 5.1. An optional fast mode costs $16 and $80 per million tokens and speeds token generation by up to 2.5 times. Anthropic says outputs run up to 30% faster than version 5.1 and that Fable 5.2 uses fewer tokens on challenging tasks at medium effort, with about 55% fewer safeguard blocks than Fable 5.1.

Benchmarks published by Anthropic: agentic coding FrontierCode 2.0 at 49.1% (Xhigh effort), above Fable 5.1's 32.4%, Opus 5.5's 45.5%, GPT-6 Astra's 42.8% and Gemini 4 Carbon's 41.1%; FrontierSWE v2 at 69.9%, above Opus 5.5's 62.3% and GPT-6 Astra's 65.5%; agentic scientific research Terminal-Bench-Science at 74.6% (75.3% for Mythos 5.2); CVE rediscovery at 89.6% for Mythos 5.2, above Opus 5.5's 81.3% without safeguards; and Humanity's Last Exam with tools at 74.2%, above Opus 5.5's 67.7%.

Anthropic says a team of 1,000 Fable 5.2 agents formalized the proof of Fermat's Last Theorem in Lean in 30 hours, against 11 days for its first attempt; the model solved two open problems in black-hole physics (published as preprints awaiting independent scrutiny), designed three working electronic devices, and accurately modeled 28 of 34 protein capsid structures attempted. Fable 5.2 is available immediately on all platforms — Claude, the Claude Platform as claude-fable-5-2, Amazon Web Services, Google Cloud and Microsoft Azure.

Supply-Chain & Market Read-Through: A 20% price cut combined with 30% faster outputs sharpens frontier-model competition on both price and performance: OpenAI's GPT-6 Astra lists at $10 per million input tokens and $50 per million output tokens, so Anthropic's new rates undercut it by 20% and put downward pressure on the API pricing band. Anthropic says Fable 5.2 at high effort matches Opus 5.5 at maximum effort for roughly half the cost per task, making inference cost per task the new competitive axis. Mythos 5.2's cybersecurity capability — an 89.6% CVE rediscovery rate — raises the bar for security-research tooling and entrenches the "restricted model plus verification program" pattern across the industry. On the compute side, 1,000-agent coordinated workloads point to rising inference intensity, supporting demand for AI accelerators and HBM; Fable 5.2 runs on both Nvidia GPUs and Google TPUs, so the beneficiaries of model competition sit more in the compute and memory supply chain than in any single chip vendor. The timing matters too: the launch lands the day before Anthropic's IPO roadshow window opens on November 10, giving the roadshow a fresh product narrative.

Next Catalysts & Watchpoints:

• Anthropic IPO roadshow: earliest start November 10; a price range is still to be filed in a further S-1/A amendment, with earliest pricing on November 13 and earliest first trade on November 16.
• Independent scrutiny of the black-hole physics preprints and peer review of the 1,000-agent formalization result.
• Whether OpenAI confirms GPT-6.2 Astra/Sol, still unannounced as of this check.

Sources:

• Anthropic: Introducing Claude Fable 5.2 and Claude Mythos 5.2 (Published: 2026-11-09 14:38 ET / 19:38 UTC)
• MacRumors: Anthropic Launches Claude Fable 5.2 and Mythos 5.2 (Published: 2026-11-09 14:38 ET / 19:38 UTC)
• Securities.io: Anthropic Launches Claude Fable and Mythos 5.2 at Lower API Prices

Information only, not investment advice.