Cursor Launches Composer 2 to Compete with OpenAI and Anthropic

Cursor, a San Francisco-based artificial intelligence startup, has launched Composer 2, a new AI model designed specifically for software development. This model aims to rival established players like OpenAI and Anthropic PBC by offering enhanced capabilities for complex coding tasks. Composer 2 is positioned as an AI agent that autonomously handles intricate programming challenges, marking a significant advancement from its predecessor, Composer 1.5.

BloombergCursor moves to build its own AI coding model to challenge OpenAI and AnthropicBenzingaVentureBeat

Composer 2 introduces a more efficient coding model, boasting a drastic reduction in costs compared to earlier versions. The standard version is priced at $0.50/$2.50 per million input/output tokens, while the faster variant, Composer 2 Fast, costs $1.50/$7.50 per million tokens. This pricing represents an approximately 86% decrease from Composer 1.5, which was priced at $3.50 and $17.50 for input and output tokens, respectively. The model is designed to support long-horizon coding tasks, with a context window of 200,000 tokens, allowing it to manage complex workflows that involve multiple file edits and command executions.

VentureBeatCursor moves to build its own AI coding model to challenge OpenAI and AnthropicBloomberg

Cursor has reported significant performance improvements with Composer 2, achieving scores of 61.3 on CursorBench, 61.7 on Terminal-Bench 2.0, and 73.7 on SWE-bench Multilingual. These benchmarks indicate a clear advancement over Composer 1.5, which scored 44.2, 47.9, and 65.9, respectively. While Composer 2 does not claim universal leadership—GPT-5.4 still leads in some benchmarks—it is positioned as a competitive option that balances performance with cost efficiency.

VentureBeat

The launch of Composer 2 reflects Cursor's strategic shift towards developing proprietary models to reduce reliance on third-party AI systems. By focusing on a specialized coding model, Cursor aims to enhance speed and efficiency in coding workflows, addressing a key challenge of dependency on external providers. This move is crucial as competition intensifies in the AI coding space, with tools like OpenAI’s Codex and Anthropic’s Claude Code gaining traction among developers.

Cursor moves to build its own AI coding model to challenge OpenAI and AnthropicVentureBeatBloomberg

Despite the advancements, Cursor faces challenges in proving the value of its integrated platform. Some users have expressed frustration with Cursor's pricing and context loss, leading them to consider alternatives like Anthropic’s Claude Code. This feedback highlights the need for Cursor to demonstrate that its in-house models and integrated features provide sufficient value to justify its position as an intermediary between developers and leading AI coding products.

VentureBeat

The introduction of Composer 2 is a pivotal moment for Cursor as it seeks to establish itself as a standalone player in the AI coding market. By focusing on specialized, efficient models, Cursor aims to reshape its business model, reduce operational costs, and enhance margins. As the AI-assisted programming landscape evolves, Cursor's strategy underscores a broader trend towards domain-specific excellence, positioning itself to compete effectively against larger firms investing heavily in developer-focused tools.

Cursor moves to build its own AI coding model to challenge OpenAI and AnthropicBloombergVentureBeatBenzinga