Based on 1 review
Anthropic's latest and most capable model with 200K standard context and 1M tokens available in beta, plus 128K max output tokens. Achieves 76% long-context retrieval accuracy at 1M tokens, roughly 4x better than Sonnet. The current state of the art for agentic coding, research, and complex software engineering tasks.
Released
February 5, 2026
Parameters
Unknown
Context
200K
Pricing
Paid
Last updated: March 15, 2026
Benchmark scores may vary based on evaluation methodology and conditions.