“In an internal agentic coding evaluation, Claude 3.5 Sonnet solved 64% of problems, outperforming Claude 3 Opus which solved 38%.”
Anthropic’s mid-tier model now beats its own flagship at $3 per million input tokens and $15 per million output. The headline coding number comes from an internal evaluation nobody outside can check. The real casualty is Claude 3 Opus. Anyone still paying flagship prices gets half the speed and the lower score.