Claude 4 Launches, Surpassing AI Competitors in Benchmarks
- Claude Opus 4 scored 72.5% on SWE-bench Verified, outperforming GPT-4.1’s 54.6% and Gemini 2.5 Pro’s 63.2%.
- The models can autonomously code for up to seven hours and handle nearly one million token context windows.
- Anthropic charges $75 per million output tokens for Claude Opus 4, significantly higher than open-source alternatives.
- The company reported $2 billion in annualized revenue during Q1, doubling from previous periods.
Claude Opus 4 has set a new standard with a high score on coding benchmarks and extended autonomous capabilities, though at a premium cost compared to competitors.
Source (2.6)https://decrypt.co/321648/claude-4-obliterates-ai-rivals-budgets-too?rand=52368
