Benchmark
Vals Router Index
Scores are percentages.
Benchmark sourceScores from Vals AI.
- Models measured
- 24
- Scores recorded
- 24
- Sources
- 1
- IQ weight
- 3.08
Score distribution
- 56.3%
- Mean
- 57.8%
- Median
- 10.1%
- Std dev
- 50.9%
- 10th percentile
- 66.8%
- 90th percentile
- 27.6%
- Min
This benchmark has a weight of 3.08 for chat models, measured over 24 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Gemini 4 Argon· Vals AI | 69.2% |
| 2 | Claude Sonnet 5.5· Vals AI | 69% |
| 3 | Muse Wonder Full· Vals AI | 66.8% |
| 4 | GPT-6 Astra· Vals AI | 65% |
| 5 | Muse Wonder· Vals AI | 64.9% |
| 6 | Claude Opus 5.5· Vals AI | 64.2% |
| 7 | GPT-6.1 Sol· Vals AI | 63.9% |
| 8 | DeepSeek V4.1 Flash 1017· Vals AI | 61.3% |
| 9 | GLM 5.3· Vals AI | 59.7% |
| 10 | Claude Haiku 5.5· Vals AI | 58.7% |
| 11 | Grok 4.7· Vals AI | 58.6% |
| 12 | MiMo V2.6 Pro· Vals AI | 57.8% |
| 13 | Gemini 3.8 Flash· Vals AI | 56% |
| 14 | Jev Router· Vals AI | 55.7% |
| 15 | MiMo V2.6 Flash· Vals AI | 54.8% |
| 16 | MiniMax M3.1 Flash· Vals AI | 53.8% |
| 17 | Hy4 preview· Vals AI | 53.2% |
| 18 | DeepSeek V4 Pro 0813· Vals AI | 53% |
| 19 | Kimi K3· Vals AI | 53% |
| 20 | GPT-6 Luna· Vals AI | 52.3% |
| 21 | OpenRouter Auto Router· Vals AI | 52.2% |
| 22 | GLM 5.3 Flash· Vals AI | 50.9% |
| 23 | Inkling FP4· Vals AI | 29% |
| 24 | Inkling Small· Vals AI | 27.6% |