Terminal-Bench — Kimi K2 Thinking card; distinct tools/format axis, vintage (1.0 vs 2.x) not specified
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Claude Sonnet 4.5 | Anthropic | 51 | 1 | 2026-06-12 |
| 2 | Kimi K2 (Reasoning) | Moonshot | 47.1 | 2 | 2026-06-12 |
| 3 | GPT-5 | OpenAI | 43.8 | 1 | 2026-06-12 |
| 4 | DeepSeek-V3.2 | DeepSeek | 37.7 | 1 | 2026-06-12 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.