HMMT 2025 — No-tools configuration, distinct from heavy and w/python tool-configurations.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | GPT-5 | OpenAI | 93.3 | 1 | 2026-06-12 |
| 2 | Grok 4 | SpaceXAI | 90 | 1 | 2026-06-12 |
| 3 | Kimi K2 (Reasoning) | Moonshot | 89.4 | 2 | 2026-06-12 |
| 4 | DeepSeek-V3.2 | DeepSeek | 83.6 | 1 | 2026-06-12 |
| 5 | Claude Sonnet 4.5 | Anthropic | 74.6 | 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.