Onyx Open LLM Leaderboard: leaderboard

Onyx Open LLM Leaderboard — curated open-model comparison table; headline MMLU-Pro column.

Metric: MMLU-Pro. Source: www.onyx.app. Saturation forecast: Estimated already saturated. 19 models tracked.

Top models

#ModelScoreOverall rank
1GPT-OSS-120B90#341
2Qwen 3.5 397B A17B87.8#146
3DeepSeek V4 Pro87.5#102
4Nemotron 3 Ultra86.8#235
5DeepSeek V4 Flash86.2#137
6Qwen 3.6 27B86.2#170
7Step 3.5 Flash85.8#270
8Gemma 4 31B85.2#200
9Qwen 3.6 35B A3B85.2#226
10DeepSeek V3.285#205
11MiMo-V2-Flash84.9#372
12GLM-4.784.3#192
13DeepSeek R184#251
14Nemotron 3 Super83.7#335
15Llama 4 Maverick80.5#464

No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=onyx-open-llm-leaderboard · How It Works · Data refreshed daily, snapshot 2026-10-20.