BBH from the HF Open LLM Leaderboard v2 — its own NORMALIZED scale (per-subtask baselines, not invertible from the aggregate; verified). A separate measurement series by design. Unpooled.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Qwen 14B B100 | Alibaba | 49.8 | 1 | 2026-07-04 |
| 2 | Yi 1.5 34B Chat 16K | 01.AI | 44.5 | 1 | 2026-07-04 |
| 3 | Yi 1.5 34B Chat | 01.AI | 44.3 | 1 | 2026-07-04 |
| 4 | Yi 1.5 34B 32K | 01.AI | 43.4 | 1 | 2026-07-04 |
| 5 | Yi 1.5 34B | 01.AI | 42.7 | 1 | 2026-07-04 |
| 6 | Yi 1.5 9B Chat | 01.AI | 37 | 1 | 2026-07-04 |
| 7 | Marco-o1 | AIDC-AI | 34.8 | 1 | 2026-07-04 |
| 8 | Yi 1.5 9B Chat 16K | 01.AI | 31.5 | 1 | 2026-07-04 |
| 9 | Yi 1.5 9B | 01.AI | 30.5 | 1 | 2026-07-04 |
| 10 | Qwen2.5 7B Test Novelist | Alibaba | 30.4 | 1 | 2026-07-04 |
| 11 | Yi 1.5 9B 32K | 01.AI | 28.9 | 1 | 2026-07-04 |
| 12 | Yi 9B | 01.AI | 27.6 | 1 | 2026-07-04 |
| 13 | Yi 9B 200K | 01.AI | 26.5 | 1 | 2026-07-04 |
| 14 | Yi Coder 9B Chat | 01.AI | 25.9 | 1 | 2026-07-04 |
| 15 | smartllama3.1-8B-001 | Meta | 24.9 | 1 | 2026-07-04 |
| 16 | Yi 1.5 6B Chat | 01.AI | 23.7 | 1 | 2026-07-04 |
| 17 | Yi 1.5 6B | 01.AI | 22 | 1 | 2026-07-04 |
| 18 | Llama Squared 8B | Meta | 21.3 | 1 | 2026-07-04 |
| 19 | Qwen2.5 1.5B Continuous Learnt | Alibaba | 19.8 | 1 | 2026-07-04 |
| 20 | NuminaMath-7B-CoT | AI-MO | 19.2 | 1 | 2026-07-04 |
| 21 | Llama 3 Instruct 8B | Meta | 18.4 | 1 | 2026-07-04 |
| 22 | NuminaMath-7B-TIR | AI-MO | 16.9 | 1 | 2026-07-04 |
| 23 | Llama 3.1 8B Squareroot | Meta | 8.6 | 1 | 2026-07-04 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.