Anthropic Ships Claude Opus 4.7: 87.6% SWE-Bench Verified, New Tokenizer, 1M Context

Anthropic / VentureBeat·high signal

Anthropic released Claude Opus 4.7 on April 17 with 87.6% on SWE-bench Verified, 64.3% on SWE-bench Pro, and 69.4% on Terminal-Bench 2.0 — extending the lead in agentic coding. The model uses an updated tokenizer that increases input token counts by 1.0-1.35x and supports images up to 2,576px on the long edge. Pricing held at $5/$25 per million tokens, available on Bedrock, Vertex AI, and Foundry. Anthropic also launched a Cyber Verification Program for security researchers to use the model with fewer restrictions, while automated safeguards block prohibited cyber activity.