Benchmark
Vals Multimodal Index
Scores are percentages.
Benchmark sourceScores from Vals AI.
- Models measured
- 33
- Scores recorded
- 33
- Sources
- 1
- IQ weight
- 1.98
Score distribution
- 58.9%
- Mean
- 60.6%
- Median
- 11%
- Std dev
- 42.9%
- 10th percentile
- 72.6%
- 90th percentile
- 34.8%
- Min
This benchmark has a weight of 1.98 for chat models, measured over 33 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 | Claude Fable 5· Vals AI | 74.2% |
| 2 | Claude Opus 5· Vals AI | 73.9% |
| 3 | Kimi K3· Vals AI | 73.4% |
| 4 | GPT-5.6 Sol· Vals AI | 72.6% |
| 5 | Claude Opus 4.8· Vals AI | 70.9% |
| 6 | Muse Spark 1.2· Vals AI | 69.8% |
| 7 | GPT-5.6 Luna· Vals AI | 69.1% |
| 8 | Claude Sonnet 5· Vals AI | 68.8% |
| 9 | GPT 5.5· Vals AI | 68.1% |
| 10 | Claude Opus 4.7· Vals AI | 67.4% |
| 11 | Muse Spark 1.1· Vals AI | 66.7% |
| 12 | Qwen3.8 Max· Vals AI | 65.4% |
| 13 | Gemini 3.6 Flash· Vals AI | 65.1% |
| 14 | GPT-5.6 Terra· Vals AI | 65.1% |
| 15 | Grok 4.5· Vals AI | 63.4% |
| 16 | Gemini 3.5 Flash· Vals AI | 62.9% |
| 17 | Claude Sonnet 4.6· Vals AI | 60.6% |
| 18 | MiniMax M3· Vals AI | 60% |
| 19 | Kimi K2.6· Vals AI | 56.4% |
| 20 | Gemini 3.1 Pro Preview· Vals AI | 56.1% |
| 21 | GPT-5.4 mini· Vals AI | 54.2% |
| 22 | Qwen3.7 Plus· Vals AI | 53.9% |
| 23 | MiMo V2.5· Vals AI | 52.8% |
| 24 | Gemini 3 Flash· Vals AI | 52.2% |
| 25 | Qwen3.6 Plus· Vals AI | 51.5% |
| 26 | Inkling Small· Vals AI | 50.1% |
| 27 | Inkling FP4· Vals AI | 49.4% |
| 28 | GPT 5.4 Nano· Vals AI | 47.6% |
| 29 | Grok 4.3· Vals AI | 43.3% |
| 30 | Claude Haiku 4.5thinking· Vals AI | 42.9% |
| 31 | Gemini 3.1 Flash Lite· Vals AI | 41.4% |
| 32 | Grok 4.20· Vals AI | 39.1% |
| 33 | Mistral Medium 3.5· Vals AI | 34.8% |