Models
Last updated
As of 2026-10-14, among models on this board with a published CursorBench 4.0 score, Claude Opus 5.5 leads at 57.8%, and Gemini 3.8 Flash has the lowest listed standard input price at $0.75 per 1M tokens.
Public list prices per 1M tokens, context windows, and published SWE-bench Verified and CursorBench scores for current frontier coding models.
Prices are standard public list rates per 1 million tokens. A blank cell means that figure is not on the cited source. SWE-bench Verified is filled only from the official leaderboard. CursorBench figures are CursorBench 4.0; the note names the effort level.
This compilation is licensed under CC BY 4.0. The prices and scores remain the publishers'. CC BY 4.0 · JSON · Markdown
| Model | Input / 1M | Output / 1M | Context | SWE-bench | CursorBench |
|---|---|---|---|---|---|
Claude Opus 5.5Anthropic · claude-opus-5-5 | $4 | $20 | 1M | Not published | 57.8% |
Claude Sonnet 5.5Anthropic · claude-sonnet-5-5 | $2 | $10 | 1M | Not published | 55.5% |
Gemini 4 ArgonGoogle · gemini-4-argon | $2 | $10 | 1M | Not published | 52.9% |
Claude Fable 5.1Anthropic · claude-fable-5-1 | $10 | $50 | 1M | Not published | 51.8% |
Grok 4.7xAI · grok-4.7 | $2 | $6 | 500K | Not published | 46.3% |
GLM-5.3Z.ai · glm-5.3 | $1.4 | $4.4 | 1M | Not published | 42.6% |
DeepSeek V4 ProDeepSeek · deepseek-v4-pro | $1.32 | $3.96 | 1M | Not published | Not published |
GPT-5.6 SolOpenAI · gpt-5.6-sol | $4 | $20 | 1.05M | Not published | 41.7% |
GPT-5.6 TerraOpenAI · gpt-5.6-terra | $2 | $12 | 1.05M | Not published | 41.3% |
Gemini 3.8 FlashGoogle · gemini-3.8-flash | $0.75 | $3.75 | Not published | Not published | 39.6% |
Composer 2.5Cursor | Not published | Not published | Not published | Not published | 27.7% |
Sources and notes
- Claude Opus 5.5
CursorBench 4.0, Max effort.
- Claude Sonnet 5.5
CursorBench 4.0, Max effort.
- Gemini 4 Argon
Standard paid tier through November 30, 2026. The listed rate becomes $4 input and $20 output on December 1, 2026.
CursorBench 4.0, High effort.
- Claude Fable 5.1
CursorBench 4.0, Max effort.
- Grok 4.7
Price for prompts under 200K tokens. A prompt of 200K tokens or more is billed at $4 input and $12 output per 1M tokens for the whole request.
CursorBench 4.0, Extra High effort.
- GLM-5.3
CursorBench 4.0, Max effort.
- DeepSeek V4 Pro
Cache-miss price during peak hours (01:00–04:00 and 06:00–10:00 UTC, Monday–Friday, excluding Chinese public holidays). Off-peak is half. The docs list the model version as DeepSeek-V4-Pro-0813.
- GPT-5.6 Sol
Prompts over 272K input tokens are billed at 2× input and 1.5× output for the whole request. OpenAI lists this rate as promotional at least through November 21, 2026.
CursorBench 4.0, Max effort.
- GPT-5.6 Terra
Prompts over 272K input tokens are billed at 2× input and 1.5× output for the whole request.
CursorBench 4.0, Max effort.
- Gemini 3.8 Flash
Standard paid tier through December 31, 2026. The listed rate becomes $1.50 input and $7.50 output on January 1, 2027.
CursorBench 4.0, High effort.
- Composer 2.5
CursorBench 4.0. The published row does not name an effort level.