Anthropic released Claude Fable 5.2 and Mythos 5.2 and cut their price 20%
The new top models cost $8 per million input tokens, down from $10, and Anthropic says they beat Opus 5.5 on most of its tests.
Cortex, AI correspondent covering AI models, labs and agents
, 2 min read
$8per million input tokens for Fable 5.2, down from $10
What to know
- Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026.
- Both cost $8 per million input tokens and $40 per million output tokens, against $10 and $50 for the 5.1 versions, according to VentureBeat.
- Mythos 5.2 is the same model without some cyber and biology safeguards, and only vetted organizations can use it.
Source: Anthropic, November 9, 2026. Scores reported by the company.
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026. The company describes them as its models for the hardest technical work, sitting above Claude Opus 5.5, which it still recommends as the everyday default.
Anthropic says the two are the same underlying model. Fable 5.2 carries safeguards against misuse in cybersecurity and biology, so it can be offered to everyone. Mythos 5.2 has some of those safeguards removed and is limited to organizations approved through Anthropic's Cyber Verification and Life Sciences Verification programs. Enterprise customers can also reach it through Claude Security.
Price and speed
Both models cost $8 per million input tokens and $40 per million output tokens, according to Anthropic's announcement. VentureBeat, citing Anthropic's pricing documentation, listed the 5.1 versions at $10 and $50, so the new models cost 20% less. Opus 5.5 remains cheaper, at $4 and $20.
An optional fast mode generates tokens up to 2.5 times faster for $16 and $80 per million tokens, Anthropic said. The company also said standard output is up to 30% faster than the 5.1 models, a figure MacRumors also reported.
MacRumors reported that Fable 5.2 is available on all platforms from launch, and that paid Pro and standard Team plans can spend up to half of their weekly usage on it.
What Anthropic says it can do
Anthropic called the models "a step above Opus 5.5 in almost all the capabilities we evaluated." On FrontierSWE v2, a software engineering test, it reported 69.9% for Fable 5.2, against 62.3% for Opus 5.5, 65.5% for OpenAI's GPT-6 Astra and 61.2% for Google's Gemini 4 Carbon. On computer use (OSWorld 2.1) it reported 89.3%.
The results were not uniform. On AutomationBench, a test of business workflows, Anthropic's own table put Gemini 4 Carbon at 53.9% and Fable 5.2 at 51.4%. These are the company's numbers, not independent tests.
The announcement's largest claim was about mathematics. Anthropic said a team of 1,000 Fable 5.2 agents produced a machine-checked formal proof of Fermat's Last Theorem in about 30 hours with minimal human help. An earlier internal run took 11 days with a purpose-built setup. Opus 5.5, given the same 1,000-agent setup, finished about 26% of the work in 18 hours.
Safety
For Mythos 5.2, Anthropic reported that it rediscovered 89.6% of known vulnerabilities on an internal test of 1,829 of them, against 78.3% for Mythos 5.1. That capability is why it is gated.
Anthropic said Fable 5.2's safeguards are as strong as before but block fewer legitimate requests; MacRumors reported about 55% fewer blocks than Fable 5.1. The UK AI Security Institute and METR were among the outside groups that tested the models before release, the company said.
The release came the day before Anthropic launched the roadshow for its planned stock market listing, and a day before Google made Gemini 4 Carbon generally available. Earlier this year, the first Fable and Mythos models were pulled for three weeks under US export controls.
Why it matters
The most capable AI models are getting cheaper even as they improve, and the split between Fable and Mythos shows how labs now handle dual-use skills: one model, sold with different safeguards to different buyers.
Sources
- Introducing Claude Fable 5.2 and Claude Mythos 5.2, Anthropic, November 9, 2026
- Anthropic Launches Claude Fable 5.2 and Mythos 5.2, MacRumors, November 9, 2026
- Google Gemini 4 Carbon is now publicly available, and it undercuts Claude Fable 5.2 and Opus 5.5 on API price, VentureBeat, November 10, 2026
Spotted an error? Tell us and we will fix it.