Nejumi 4 - ALT Average: leaderboard
The unweighted mean of Nejumi 4's six alignment categories: controllability, ethics and morality, toxicity, bias, truthfulness and robustness. It is an index computed from those category scores rather than a task the model runs, and it is reported on a 0 to 100 scale where higher is better. Source: github.com/wandb/llm-leaderboard.
Metric: Score (%). Source: nejumi.ai. Saturation forecast: Estimated already saturated. 139 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.6 Max Preview | 91.81 | #85 |
| 2 | Qwen3.8 2.4T A95B | 91.71 | #55 |
| 3 | Qwen 3.5 122B A10B | 90.78 | #173 |
| 4 | Claude Fable 5 | 90.71 | #16 |
| 5 | Qwen 3.5 397B A17B | 90.68 | #146 |
| 6 | Qwen 3.6 35B A3B | 90.2 | #226 |
| 7 | Kimi K3 | 90.14 | #30 |
| 8 | GPT-5.4 (2026-03-05) | 89.93 | #59 |
| 9 | Muse Spark Open | 89.9 | #54 |
| 10 | Claude Opus 4.6 | 89.87 | #63 |
| 11 | Claude Opus 4.1 | 89.81 | #140 |
| 12 | Solar Pro 4 | 89.57 | #215 |
| 13 | GPT-5.2 | 89.54 | #110 |
| 14 | Claude Opus 4.7 | 89.51 | #49 |
| 15 | Qwen 3.5 27B | 89.47 | #187 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=nejumi-4-alt-average · How It Works · Data refreshed daily, snapshot 2026-10-20.