MCPMark (Verified subset) — Likely Verified subset of mcpmark, but diff source/naming, no raw overlap -- kept separate.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Kimi K3 | Moonshot | 94.5 | 1 | 2026-07-27 |
| 2 | GPT-5.5 | OpenAI | 92.9 | 2 | 2026-07-27 |
| 3 | GPT-5.6 Sol | OpenAI | 92.9 | 1 | 2026-07-27 |
| 4 | Claude Fable 5 | Anthropic | 87.4 | 1 | 2026-07-27 |
| 5 | Kimi K2.7 Code | Moonshot | 81.1 | 1 | 2026-06-14 |
| 6 | Claude Opus 4.8 | Anthropic | 76.4 | 2 | 2026-07-27 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.