Benchmark
Vals SAGE
Scores are percentages.
Benchmark sourceScores from Vals AI.
- Models measured
- 29
- Scores recorded
- 29
- Sources
- 1
- IQ weight
- 2.24
Score distribution
- 84.5%
- Mean
- 85.5%
- Median
- 6.4%
- Std dev
- 79.5%
- 10th percentile
- 89.5%
- 90th percentile
- 59.2%
- Min
This benchmark has a weight of 2.24 for chat models, measured over 29 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 Opus 5.5· Vals AI | 91.3% |
| 2 | GPT-6.1 Sol· Vals AI | 90.5% |
| 3 | Claude Haiku 5.5· Vals AI | 89.8% |
| 4 | Gemini 4 Argon· Vals AI | 89.5% |
| 5 | GLM 5.3 Flash· Vals AI | 89.3% |
| 6 | Qwen3.7 Plus· Vals AI | 89.1% |
| 7 | Grok 4.7· Vals AI | 88.7% |
| 8 | Inkling 2 Small· Vals AI | 88.6% |
| 9 | Gemini 3.8 Flash· Vals AI | 88.2% |
| 10 | Claude Sonnet 5.5· Vals AI | 88% |
| 11 | GPT-5.6 Terra· Vals AI | 87% |
| 12 | Kimi K3· Vals AI | 86.8% |
| 13 | Ember-1· Vals AI | 86.4% |
| 14 | Muse Spark Open· Vals AI | 85.8% |
| 15 | MiniMax M3· Vals AI | 85.5% |
| 16 | Qwen3.5 397B A17B· Vals AI | 85.5% |
| 17 | DeepSeek V4.1 Flash· Vals AI | 85.4% |
| 18 | MiniMax M3.1 Flash· Vals AI | 85.3% |
| 19 | Qwen3.5 Flash· Vals AI | 84.9% |
| 20 | MiMo V2.6 Pro· Vals AI | 84.4% |
| 21 | GPT-6 Luna· Vals AI | 83.4% |
| 22 | MiMo V2.6 Flash· Vals AI | 83.3% |
| 23 | Gemini 3.1 Pro Preview· Vals AI | 83.2% |
| 24 | Gemma 4 31B IT· Vals AI | 82.8% |
| 25 | Qwen3.8 Max· Vals AI | 81.8% |
| 26 | Gemini 3.5 Flash Lite· Vals AI | 79.5% |
| 27 | Grok 4.1 Fast· Vals AI | 74.5% |
| 28 | Mistral Medium 3.5· Vals AI | 72% |
| 29 | Mistral Small2603· Vals AI | 59.2% |