Gate News, March 20 — AI programming tool Cursor announced the release of its third-generation in-house developed programming model, Composer 2. The standard version is priced at $0.50 per million tokens for input and $2.50 for output, approximately 86% cheaper than the previous Composer 1.5. A fast version is also available, costing $1.50 per million tokens for input and $7.50 for output, and is set as the default option.
Benchmark tests published by Cursor show that Composer 2 scored 61.7% on the Terminal-Bench 2.0 programming agent benchmark, surpassing Opus 4.6’s 58.0%, but falling short of GPT-5.4’s 75.1%. On the multilingual SWE-bench software engineering benchmark, it scored 73.7% (up from 65.9% in the previous version), and on Cursor’s proprietary CursorBench, it scored 61.3% (up from 44.2%).
Cursor stated that the performance improvements come from continued pretraining of the base model and reinforcement learning to train the model to complete long-cycle programming tasks requiring hundreds of steps. This model is only available within Cursor and is not offered as an independent API. Cursor’s parent company, Anysphere, is currently valued at $29.3 billion.
Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to
Disclaimer.
