Key Points
- Anthropic launched Claude Mythos 5.2 and Claude Fable 5.2, cutting standard token fees by 20% while boosting speed and introducing a new 2.5x Fast Mode.
- The flagship model scored 65 on Artificial Analysis's intelligence benchmark, overtaking Google's Gemini 4 Carbon and OpenAI's GPT-6 Astra for the top spot.
- The release comes as a decisive strategic move directly preceding Anthropic's IPO scheduled for this month, despite current market volatility.
Summary Anthropic has unveiled its next-generation artificial intelligence flagship models, 'Claude Mythos 5.2' and 'Claude Fable 5.2,' in a strategic move ahead of its upcoming initial public offering (IPO). Aiming to solidify its market dominance against fierce rivals in the AI race, the company announced major gains in benchmark performance, generation speed, and pricing efficiency.
According to performance figures released with the launch, the newly released 5.2 iterations boast up to a 30% faster output generation speed compared to version 5.1. In addition, Anthropic rolled out a high-throughput 'Fast Mode' that increases inference speed by up to 2.5 times while preserving full intelligence capabilities. Notably, the standard pricing per million tokens was slashed by 20% to between $8 and $40, while the Fast Mode is priced from $16 to $80 per million tokens.
In comparative benchmark evaluations covering coding, cybersecurity, and scientific exploration, Anthropic's new models surpassed existing competitors, including OpenAI’s GPT-6 Astra and Google’s Gemini 4 Carbon. Artificial Analysis awarded the model an intelligence index (AAII) score of 65—eclipsing Gemini 4 Carbon's previous top mark of 60 points by a wide margin, with Claude Opus 5.5 (59 points) and GPT-6 Astra (57 points) trailing behind.
Anthropic highlighted remarkable scientific problem-solving capabilities, reporting that Mythos and Fable 5.2 resolved two unsolved theoretical physics problems and formalized a proof of Fermat's Last Theorem into code within 30 hours. On safety and practical usability, the company reduced unnecessary false-positive refusals by 55% while maintaining strict biological and cyber defense safeguards. Despite headwinds from broader AI stock sell-offs and industry turbulence, Anthropic reportedly plans to complete its IPO within the month.
Sponsored · Ad