Cursor Launches Composer 2, Rivaling Anthropic and OpenAI at a Fraction of the Cost
Summary
Cursor launches Composer 2, a powerful new coding model that matches Anthropic's Claude Opus 4.6 on key benchmarks while costing 10 times less at just $0.50 per million input tokens, signaling a major shift in the AI coding tools market.
Key Points
- Cursor launches Composer 2, its most capable coding model yet, positioning it as a low-cost alternative to competitors like Anthropic's Claude Code and OpenAI's Codex.
- On the Terminal-Bench 2.0 benchmark, Composer 2 scores 59.8% — comparable to Anthropic's Opus 4.6 at 58.0% — while costing significantly less at $0.50/M input tokens versus Opus 4.6's $5/M input tokens.
- Cursor differentiates itself beyond just model performance by offering a full development environment that supports multiple AI models, long-running agents, code review, a plugin marketplace, and enterprise security features.