Onyx Open LLM Leaderboard: leaderboard

Metric: MMLU-Pro. Source: www.onyx.app. 19 models tracked.

Top models

#ModelScore
1GPT-OSS-120B90
2Qwen 3.5 397B A17B87.8
3DeepSeek V4 Pro87.5
4Nemotron 3 Ultra86.8
5DeepSeek V4 Flash86.2
6Qwen 3.6 27B86.2
7Step 3.5 Flash85.8
8Gemma 4 31B85.2
9Qwen 3.6 35B A3B85.2
10DeepSeek V3.285
11MiMo-V2-Flash84.9
12GLM-4.784.3
13DeepSeek R184
14Nemotron 3 Super83.7
15Llama 4 Maverick80.5

Interactive version: theaggregate.ai/benchmark?slug=onyx-open-llm-leaderboard · How It Works · Data refreshed daily, snapshot 2026-09-05.