ResearchRubrics — Rubric-graded deep-research-agent benchmark.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Kimi K3 | Moonshot | 76.2 | 1 | 2026-07-27 |
| 2 | GPT-5.6 Sol | OpenAI | 73.8 | 1 | 2026-07-27 |
| 3 | Claude Opus 4.8 | Anthropic | 73.5 | 1 | 2026-07-27 |
| 4 | GLM-5.2 Full Open Source | Z.ai | 71.1 | 1 | 2026-07-27 |
| 5 | Step 3.5 Flash | StepFun | 65.3 | 3 | 2026-05-30 |
| 6 | GPT-5.5 | OpenAI | 64 | 1 | 2026-07-27 |
| 7 | GLM-4.7 | Z.ai | 62 | 3 | 2026-05-30 |
| 8 | MiniMax M2.1 | MiniMax | 60.2 | 3 | 2026-05-30 |
| 9 | Kimi K2.5 | Moonshot | 59.5 | 3 | 2026-05-30 |
| 10 | Kimi K2 (Reasoning) | Moonshot | 56.2 | 3 | 2026-05-30 |
| 11 | DeepSeek-V3.2 | DeepSeek | 55.8 | 3 | 2026-05-30 |
| 12 | MiMo V2 Flash | Xiaomi | 54.3 | 3 | 2026-05-30 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.