Tag
5 articles
Browse all tagsMeta Superintelligence Labs has released Muse Wonder and Muse Wonder Full, agent-focused models that land near the top of the Artificial Analysis index. Wonder costs $4 / $20 per million tokens through the API. Wonder Full is only available through Muse Code. Here are the benchmarks, the pricing and the catches.
7 min read
Mistral has opened a public API preview of Mistral Large 4, a roughly 1-trillion-parameter mixture-of-experts model with a 1M-token context window. The weights are promised by the end of October. Here is what is confirmed, what it costs and what is still unknown.
4 min read
Anthropic's new Haiku 5.5 costs from $0.10 / $0.50 per million tokens, 90% less than Haiku 4.5, and makes huge gains on coding, computer use and knowledge work. Here is what is new, what it costs and what developers need to change.
7 min read
Anthropic's Claude Sonnet 5.5 runs over 30% faster than Sonnet 5, costs up to 30% less per task at the same per-token price, and lands within a few points of Opus 5.5 on most benchmarks. Here is what is new and what developers need to change.
7 min read
Anthropic's new Opus 5.5 matches or beats Claude Fable 5.1 on most work, runs over 30% faster than Opus 5 and costs about 40% less to run. Here is what is new, what it costs and what developers need to change.
5 min read