Composer 2

Model ID: composer-2
Context window: 200k
Max context: -
Provider: Cursor
Capabilities: AgentThinking
Speed: Fast
Cost: Low
Intelligence: Frontier

Composer 2 is Cursor's own agentic model. It reaches frontier-level coding performance with strong results on CursorBench while using tokens efficiently.

Strengths

• Strong on long-horizon tasks via self-summarization during training.
• Default fast variant for interactive sessions; standard tier further optimized for cost per token.
• Tuned for tool use, file edits, and terminal operations inside Cursor.

Tools

Composer 2 has access to all agent tools when used with Cursor including:

Semantic search

Search your indexed codebase by meaning, not exact matches.

Search files and folders

Find files by name, read directory structures, and grep for patterns.

Web

Generate search queries and fetch results from the web.

Read files

Read file contents, including images for vision-capable models.

Edit files

Suggest edits and apply them automatically.

Run shell commands

Execute terminal commands and monitor output.

Browser

Control a browser to take screenshots, test applications, and verify visual changes. See the Browser documentation.

Image generation

Generate images from text descriptions or reference images.

Ask questions

Ask clarifying questions while continuing to work in the background.

Fetch rules

Retrieve rules based on type and description.

Learn more about how tools work and tool calling fundamentals.

Pricing

A faster variant with the same intelligence is also available at $1.50/M input and $7.50/M output tokens. Fast is the default in the product and is priced lower than other fast models at similar speeds.

On individual plans, Composer draws from a standalone Composer usage pool with generous included usage. On team and enterprise plans, the API price below is charged directly. All prices are per million tokens.