Benchmark
Vals EMB
Scores are percentages.
Benchmark sourceScores from Vals AI.
- Models measured
- 73
- Scores recorded
- 73
- Sources
- 1
- IQ weight
- 3.14
Score distribution
- 54.4%
- Mean
- 58.1%
- Median
- 17.7%
- Std dev
- 26%
- 10th percentile
- 72.2%
- 90th percentile
- 8.6%
- Min
This benchmark has a weight of 3.14 for chat models, measured over 73 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 |
|---|---|---|
| 51 | Grok 4.5· Vals AI | 52.9% |
| 52 | DeepSeek V4 Pro 0813· Vals AI | 52.8% |
| 53 | Gemini 3.1 Pro Preview· Vals AI | 52.6% |
| 54 | DeepSeek V4 Pro· Vals AI | 51.6% |
| 55 | Qwen3.7 Plus· Vals AI | 49.3% |
| 56 | MiniMax M3· Vals AI | 47.8% |
| 57 | GPT-5.4 mini· Vals AI | 45.4% |
| 58 | GPT 5.4 Nano· Vals AI | 44.8% |
| 59 | Gemini 3.5 Flash Lite· Vals AI | 43.2% |
| 60 | Inkling FP4· Vals AI | 38.4% |
| 61 | Qwen3.6 Plus· Vals AI | 32.9% |
| 62 | Inkling Small· Vals AI | 31.8% |
| 63 | Nemotron 3 Ultra· Vals AI | 31.7% |
| 64 | Kimi K2.5thinking· Vals AI | 28.5% |
| 65 | MiniMax M2.7· Vals AI | 28.5% |
| 66 | Gemini 3 Flash· Vals AI | 26% |
| 67 | Claude Haiku 4.5thinking· Vals AI | 22.7% |
| 68 | Grok 4.3· Vals AI | 18.4% |
| 69 | Nemotron 3.5 Lightning 30B A3B· Vals AI | 17% |
| 70 | Grok 4.20· Vals AI | 11.7% |
| 71 | Mistral Medium 3.5· Vals AI | 11.2% |
| 72 | Mercury 2.5· Vals AI | 8.8% |
| 73 | Gemini 3.1 Flash Lite· Vals AI | 8.6% |
Showing 51–73 of 73 results.