MBPP (Mostly Basic Python Problems) — Base MBPP, distinct from the EvalPlus-augmented MBPP+ variants and the Sanitized subset.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Kimi K2 Base | Moonshot | 89 | 2 | 2026-06-05 |
| 2 | Granite 4.1 8B | IBM | 87.3 | 1 | 2026-06-15 |
| 3 | Granite 4.1 30B | IBM | 85.5 | 1 | 2026-06-15 |
| 4 | Granite 4.1 30B Base | IBM | 83.6 | 2 | 2026-06-15 |
| 5 | Stable-DiffCoder-8B-Base | ByteDance | 83.6 | 1 | 2026-06-05 |
| 6 | Yi Coder 9B Chat | 01.AI | 82 | 1 | 2026-06-05 |
| 7 | Nemotron 3 Super 120B A12B | NVIDIA | 81.7 | 1 | 2026-08-11 |
| 8 | GLM-4.5-Base | Z.ai | 81.6 | 2 | 2026-06-05 |
| 9 | Granite 4.1 3B Base | IBM | 81.5 | 1 | 2026-06-15 |
| 10 | MiMo V2 Flash Base | Xiaomi | 81 | 2 | 2026-06-05 |
| 11 | Seed Coder 8B Instruct | ByteDance | 79.6 | 1 | 2026-06-05 |
| 12 | Step3.5 Flash Base | StepFun | 79.4 | 2 | 2026-06-05 |
| 13 | Qwen2.5 Coder Instruct 7B | Alibaba | 79.4 | 1 | 2026-06-05 |
| 14 | Qwen2.5 3B Base | Alibaba | 79.1 | 1 | 2026-05-03 |
| 15 | Nemotron 3.5 Lightning | NVIDIA | 78.6 | 1 | 2026-08-11 |
| 16 | NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16 | NVIDIA | 78.6 | 1 | 2026-08-24 |
| 17 | Qwen3 8B | Alibaba | 77 | 1 | 2026-06-05 |
| 18 | DeepSeek-V3.2-Exp-Base | DeepSeek | 75.6 | 2 | 2026-06-05 |
| 19 | DeepSeek-V3 | DeepSeek | 75.4 | 1 | 2026-05-01 |
| 20 | DeepSeek-Coder-6.7B-Instruct | DeepSeek | 74.9 | 1 | 2026-06-05 |
| 21 | DeepSeek-V3.1-Base | DeepSeek | 74.6 | 2 | 2026-06-05 |
| 22 | NVIDIA Nemotron 3 Nano 30B A3B | NVIDIA | 73.8 | 1 | 2026-08-11 |
| 23 | Llama 3.1 Instruct 405B | Meta | 73.4 | 2 | 2026-05-31 |
| 24 | Qwen2.5 Instruct 72B | Alibaba | 72.6 | 2 | 2026-08-23 |
| 25 | Hunyuan Large | Tencent | 72.6 | 1 | 2026-05-31 |
| 26 | Qwen3.5 35B A3B | Alibaba | 70.8 | 1 | 2026-08-11 |
| 27 | DeepSeek-Coder-6.7B-Base | DeepSeek | 70.2 | 1 | 2026-06-05 |
| 28 | Llama 3.1 8B Instruct | Meta | 70.1 | 1 | 2026-06-05 |
| 29 | Llama 3.1 70B Instruct | Meta | 68.6 | 1 | 2026-05-31 |
| 30 | Gemma 4 26B A4B | 68.4 | 1 | 2026-08-11 | |
| 31 | DeepSeek-V2 | DeepSeek | 66.6 | 2 | 2026-05-31 |
| 32 | Moonlight | Moonshot | 63.8 | 2 | 2026-06-12 |
| 33 | Granite 4.1 8B Base | IBM | 63.8 | 2 | 2026-06-15 |
| 34 | Qwen2.5 3B | Alibaba | 57.1 | 2 | 2026-06-05 |
| 35 | Qwen3 4B Base Dense | Alibaba | 55.4 | 1 | 2026-05-03 |
| 36 | Llama 3.2 3B Instruct | Meta | 48.7 | 2 | 2026-06-05 |
| 37 | Qwen2.5 7B Base | Alibaba | 48.5 | 1 | 2026-05-03 |
| 38 | Gemma 3 1B Base Dense | 24.8 | 1 | 2026-05-03 | |
| 39 | LFM2.5-8B-A1B-DSpark | Liquid AI | 6.9 | 2 | 2026-08-21 |
| 40 | LFM2.5-2.6B-DSpark | Liquid AI | 4.7 | 2 | 2026-08-21 |
| 41 | Sarvam 30B | Sarvam AI | 0.9 | 1 | 2026-08-23 |
| 42 | Qwen 2.5 Coder 32B Instruct | Alibaba | 0.9 | 1 | 2026-08-23 |
| 43 | Llama 3.1 Nemotron Nano 8B V1 | NVIDIA | 0.8 | 1 | 2026-08-23 |
| 44 | Qwen2.5 VL 32B | Alibaba | 0.8 | 1 | 2026-08-23 |
| 45 | Qwen3 235B A22B | Alibaba | 0.8 | 1 | 2026-08-23 |
| 46 | Phi 3.5 MoE Instruct | Microsoft | 0.8 | 1 | 2026-08-23 |
| 47 | Qwen2 Instruct (72B) | Alibaba | 0.8 | 1 | 2026-08-23 |
| 48 | Llama 4 Maverick | Meta | 0.8 | 1 | 2026-08-23 |
| 49 | Gemma 3 27B | 0.7 | 1 | 2026-08-23 | |
| 50 | Qwen2.5 Omni 7B | Alibaba | 0.7 | 1 | 2026-08-23 |
| 51 | Gemma 3 12B | 0.7 | 1 | 2026-08-23 | |
| 52 | Phi 3.5 Mini Instruct | Microsoft | 0.7 | 1 | 2026-08-23 |
| 53 | Llama 4 Scout | Meta | 0.7 | 1 | 2026-08-23 |
| 54 | Qwen2 7B Instruct | Alibaba | 0.7 | 1 | 2026-08-23 |
| 55 | Gemma 3 4B | 0.6 | 1 | 2026-08-23 | |
| 56 | Gemma 2 27B IT | 0.6 | 1 | 2026-08-23 | |
| 57 | Gemma 2 Instruct (9B) | 0.5 | 1 | 2026-08-23 | |
| 58 | Gemma 3 1B | 0.4 | 1 | 2026-08-23 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.