On SWEBench Pro, Claude Mythos achieved a 78% score, while Opus previously scored 53% and GPT 5.4 scored 57.7%.

Claude Mythos achieved an 82% score on the terminal bench, an increase from the previous 65%.

Claude Opus 4.7 scores worse than Opus 4.6 on Simple Bench, a benchmark designed to test common sense.

Anthropic's Claude Mythos model is a much bigger, more expensive, slower, but more powerful model compared to Opus.

Claude Code is experiencing a "breakout moment" among developers, particularly with the most recent model, Opus 4.5, enabling the creation of whole apps and end-to-end tasks.

Scale AI benchmarked top models (Claude, Gemini, OpenAI) on real multi-file software engineering tasks using actual production-grade code, finding they solved only 20-30% of tasks.