MLX Benchmark V2 - MLX LM — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 90.16 |
| 2 | Qwen 3.6 Max Preview | 80.56 |
| 3 | Gemini 3 Flash (Preview) | 78.69 |
| 4 | Gemma 4 26B A4B (IT) | 77.05 |
| 5 | DeepSeek V4 Flash | 73.77 |
| 6 | Grok 4.1 Fast | 72.13 |
| 7 | DeepSeek V4 Pro | 71.93 |
| 8 | GPT-5.4 Nano | 70.49 |
| 9 | Gemini 2.5 Flash Lite (Preview 09-2025) | 68.85 |
| 10 | GPT-5 Nano | 49.18 |
| 11 | Qwen 3.6 35B A3B | 49.18 |
| 12 | GLM-5.1 | 19.67 |
| 13 | Nemotron 3 Ultra | 18.03 |
| 14 | Kimi K2.5 | 8.2 |
| 15 | Kimi K2.6 | 4.92 |
Interactive version: theaggregate.ai/benchmark?slug=mlx-benchmark-v2-mlx-lm · How the rankings work · Data refreshed daily, snapshot 2026-07-22.