MLX Benchmark V2 - MLX NN — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 97.26 |
| 2 | Gemini 3 Flash (Preview) | 90.41 |
| 3 | DeepSeek V4 Pro | 80.88 |
| 4 | GPT-5.4 Nano | 80.82 |
| 5 | DeepSeek V4 Flash | 80.82 |
| 6 | Gemini 2.5 Flash Lite (Preview 09-2025) | 76.71 |
| 7 | Gemma 4 26B A4B (IT) | 75.34 |
| 8 | Grok 4.1 Fast | 73.97 |
| 9 | Qwen 3.6 Max Preview | 73.91 |
| 10 | Qwen 3.6 35B A3B | 47.95 |
| 11 | GPT-5 Nano | 43.84 |
| 12 | GLM-5.1 | 19.18 |
| 13 | Nemotron 3 Ultra | 13.7 |
| 14 | Kimi K2.5 | 1.37 |
| 15 | Kimi K2.6 | 0 |
Interactive version: theaggregate.ai/benchmark?slug=mlx-benchmark-v2-mlx-nn · How the rankings work · Data refreshed daily, snapshot 2026-07-22.