
Composer 2 marks a significant step forward in AI coding by combining stronger performance, long-horizon task execution, and lower costs to make advanced software development more scalable and accessible. (Source: Image by RR)
Cursor Is Focusing on Balancing Intelligence, Speed, and Cost Efficiency
Cursor has released Composer 2, a next-generation coding model designed to deliver frontier-level performance while significantly improving cost efficiency. Positioned as a major upgrade over previous versions, Composer 2 combines stronger reasoning capabilities with lower token pricing. The company, according to an article in cursor.com, is aiming to hit a sweet spot between intelligence and affordability for developers building AI-powered software.
The model shows substantial gains across key coding benchmarks, including Terminal-Bench 2.0 and SWE-bench Multilingual. Composer 2 achieves a score of 61.7 on Terminal-Bench and 73.7 on SWE-bench, representing a major leap over Composer 1.5 and earlier versions. These improvements stem from Cursor’s first large-scale continued pretraining run, followed by reinforcement learning focused on long-horizon coding tasks—allowing the model to complete complex workflows that require hundreds of steps.
A key differentiator is pricing. Composer 2 is available at $0.50 per million input tokens and $2.50 per million output tokens, undercutting many competing frontier models. Cursor is also introducing a faster variant with the same intelligence but higher throughput, designed to handle real-time or high-volume use cases. The company is positioning this “fast” version as the default, signaling a push toward speed without sacrificing capability.
With Composer 2, Cursor is doubling down on its strategy of tightly integrating powerful coding models into developer workflows. By improving both performance and cost efficiency, the release reflects a broader trend in AI: moving from raw capability breakthroughs toward practical, scalable tools that can handle real-world engineering tasks at lower cost.
read more at cursor.com
Leave A Comment