MLX Benchmark V2: leaderboard
Benchmark for evaluating model knowledge of Apple's MLX machine-learning framework across implementation, debugging, and conceptual questions.
Metric: Accuracy (%). Source: huggingface.co. Source: huggingface.co. Saturation forecast: Not forecast. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 89.62 | #90 |
| 2 | Gemini 3 Flash (Preview) | 82.39 | #82 |
| 3 | Qwen 3.6 Max Preview | 80.13 | #85 |
| 4 | Gemma 4 26B A4B (IT) | 75.19 | #297 |
| 5 | GPT-5.4 Nano | 75.19 | #328 |
| 6 | Grok 4.1 Fast | 72.69 | #217 |
| 7 | DeepSeek V4 Pro | 71.84 | #102 |
| 8 | DeepSeek V4 Flash | 71.54 | #137 |
| 9 | Gemini 2.5 Flash Lite (Preview 09-2025) | 67.31 | #318 |
| 10 | Qwen 3.6 35B A3B | 52.5 | #226 |
| 11 | GPT-5 Nano | 41.92 | #425 |
| 12 | GLM-5.1 | 19.23 | #97 |
| 13 | Nemotron 3 Ultra | 16.15 | #235 |
| 14 | Kimi K2.5 | 4.81 | #144 |
| 15 | Kimi K2.6 | 3.1 | #104 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mlx-benchmark-v2 · How It Works · Data refreshed daily, snapshot 2026-10-20.