Anthropic has launched an updated mid-tier AI model, Sonnet, focusing on programming skills, following instructions, and computer usage.
This is Claude Sonnet 4.6: our most capable Sonnet model yet.
— Claude (@claudeai) February 17, 2026
It’s a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design.
It also features a 1M token context window in beta. pic.twitter.com/TDId3XUSRs
The neural network is now available by default for Free and Pro plan users.
The beta version of the LLM features a 1 million token context window—double the previous limits. Anthropic noted that the model can "accommodate entire codebases, lengthy contracts, and dozens of scientific papers in a single prompt."
The release is accompanied by new records in benchmarks such as OSWorld (computer usage) and SWE-Bench (programming tasks).
There has been steady growth in the performance of various Sonnet versions in OSWorld. Source: Anthropic.
In the ARC-AGI-2 benchmark, which assesses abstract thinking abilities, Sonnet 4.6 scored 60.4%. This score places the neural network ahead of most competitors, trailing only Opus 4.6, Gemini 3 Deep Think, and a modified version of GPT 5.2.
Sonnet 4.6's performance in various benchmarks compared to other solutions. Source: Anthropic.
Earlier in February, the startup updated its flagship model Claude Opus to version 4.6. The neural network has improved its action planning, handling of long tasks, and efficiency with large codebases.
“Performance that previously required an Opus-level model—including real office tasks with economic value—is now available in Sonnet 4.6,” the company stated in its blog.
Later, the firm reported raising $30 billion at a valuation of $380 billion. The funds will be directed towards "advanced research, product development, and infrastructure expansion."
Anthropic announced a comprehensive safety evaluation of Sonnet 4.6, stating that the new LLM "significantly outperforms" its predecessor in resilience to prompt injections.
Recall that in February, media reported on the use of the AI tool Claude in a military operation to capture Venezuelan President Nicolás Maduro.