B.O.B. Leaderboard
Alibabareleased 14 Apr 2026
Qwen3.6 35B A3B
#85 of 9513 of 46 benchmarksnot enough data
12.4
B.O.B.
Category profile
- Real use (hard to game)2×3
- Reasoning & knowledge28
- Math13
- Coding7
- Agents & real work7
- Independent indexes26
Reality check
Needs results on at least 2 real-use tests before we can compare.
Price per 1M tokens (OpenRouter)
$0.15 in · $1.00 out
Cost per task on the AA index
$0.48
Output speed (Artificial Analysis)
133 tok/s
§Receipts
Every benchmark result
“Beats” = the share of tracked models this one beat on that test. The raw score is its best reported setting.
| Benchmark | Raw | Beats | Source |
|---|---|---|---|
| Real use (hard to game) · Real use 3 | |||
| SWE-rebenchas Qwen3.6-35B-A3B | 29.2% | 3%of 34 | Nebius |
| Reasoning & knowledge · Reasoning 28 | |||
| Chess Puzzlesas qwen3.6-35b-a3b | 26.0% | 45%of 73 | Epoch AI Benchmarking Hub |
| Mystery Game Puzzlesas qwen3.6-35b-a3b_none | 22.0% | 41%of 57 | Epoch AI Benchmarking Hub |
| GPQA Diamondas qwen3.6-35b-a3b_none | 84.8% | 17%of 70 | Epoch AI Benchmarking Hub |
| CritPt (physics research)as qwen3.6-35b-a3b | 0.3% | 10%of 89 | Epoch AI Benchmarking Hub |
| Math · Math 13 | |||
| OTIS Mock AIMEas qwen3.6-35b-a3b | 86.7% | 23%of 73 | Epoch AI Benchmarking Hub |
| FrontierMath Tiers 1-3as qwen3.6-35b-a3b_none | 20.4% | 3%of 60 | Epoch AI Benchmarking Hub |
| Coding · Coding 7 | |||
| SciCodeas qwen3.6-35b-a3b | 36.6% | 9%of 86 | Epoch AI Benchmarking Hub |
| Terminal-Benchas qwen3.6-35b-a3b | 24.6% | 8%of 14 | Epoch AI Benchmarking Hub |
| WeirdMLas qwen3.6-35b-a3b | 34.5% | 4%of 52 | Epoch AI Benchmarking Hub |
| Agents & real work · Agents 7 | |||
| LMCAas qwen3.6-35b-a3b | 29.7% | 11%of 67 | Epoch AI Benchmarking Hub |
| DTBenchas qwen3.6-35b-a3b | 73.9% | 3%of 67 | Epoch AI Benchmarking Hub |
| Independent indexes · Indexes 26 | |||
| Artificial Analysis Intelligence Indexas Qwen3.6 35B A3B | 18 | 26%of 54 | Artificial Analysis |