WinoGrande — Generic/unlabeled rollup across DeepSeek/MiMo/NVIDIA cards; metric unspecified
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Hunyuan Large | Tencent | 88.7 | 1 | 2026-05-31 |
| 2 | GPT4 | OpenAI | 87.5 | 1 | 2026-08-23 |
| 3 | Llama 3.1 Instruct 405B | Meta | 86.7 | 2 | 2026-05-31 |
| 4 | DeepSeek-V2 | DeepSeek | 86.3 | 2 | 2026-05-31 |
| 5 | DeepSeek-V3.1-Base | DeepSeek | 85.9 | 4 | 2026-06-13 |
| 6 | MiMo V2.5 Pro Base | Xiaomi | 85.6 | 3 | 2026-06-05 |
| 7 | DeepSeek-V3.2-Exp-Base | DeepSeek | 85.6 | 5 | 2026-06-13 |
| 8 | MiMo V2.5 Pro | Xiaomi | 85.6 | 1 | 2026-08-23 |
| 9 | Command R+ (Apr '24) | Cohere | 85.4 | 1 | 2026-08-23 |
| 10 | Kimi K2 Base | Moonshot | 85.3 | 8 | 2026-06-13 |
| 11 | Llama 3.1 70B Instruct | Meta | 85.3 | 1 | 2026-05-31 |
| 12 | GLM-4.5-Base | Z.ai | 85.2 | 1 | 2026-06-05 |
| 13 | Qwen2 Instruct (72B) | Alibaba | 85.1 | 1 | 2026-08-23 |
| 14 | DeepSeek-V3 | DeepSeek | 84.9 | 1 | 2026-05-01 |
| 15 | MiMo V2.5 Base | Xiaomi | 84.7 | 3 | 2026-06-05 |
| 16 | Llama 3.1 Nemotron Instruct 70B | NVIDIA | 84.5 | 1 | 2026-08-23 |
| 17 | MiMo V2 Flash Base | Xiaomi | 83.8 | 4 | 2026-06-13 |
| 18 | Gemma 2 27B IT | 83.7 | 1 | 2026-08-23 | |
| 19 | Qwen2.5 Instruct 72B | Alibaba | 82.3 | 1 | 2026-05-01 |
| 20 | Mistral Large 3 675B Base 2512 | Mistral | 82.1 | 1 | 2026-06-05 |
| 21 | DeepSeek-V4-Pro-Base | DeepSeek | 81.5 | 3 | 2026-06-05 |
| 22 | Phi 3.5 MoE Instruct | Microsoft | 81.3 | 1 | 2026-08-23 |
| 23 | Qwen 2.5 Coder 32B Instruct | Alibaba | 80.8 | 1 | 2026-08-23 |
| 24 | Gemma 2 Instruct (9B) | 80.6 | 1 | 2026-08-23 | |
| 25 | NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16 | NVIDIA | 80 | 1 | 2026-08-24 |
| 26 | Nemotron 3.5 Lightning | NVIDIA | 80 | 1 | 2026-08-11 |
| 27 | DeepSeek-V4-Flash-Base | DeepSeek | 79.5 | 3 | 2026-06-05 |
| 28 | NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16 | NVIDIA | 79.3 | 1 | 2026-08-24 |
| 29 | Nemotron 3 Ultra 550B A55B Base | NVIDIA | 79.3 | 1 | 2026-06-05 |
| 30 | Qwen3.5 35B A3B | Alibaba | 79.2 | 1 | 2026-08-11 |
| 31 | NVIDIA Nemotron 3 Nano 30B A3B | NVIDIA | 79.2 | 1 | 2026-08-11 |
| 32 | Step3.5 Flash Base | StepFun | 79.1 | 2 | 2026-06-05 |
| 33 | Gemma 4 26B A4B | 79.1 | 1 | 2026-08-11 | |
| 34 | Nemotron 3 Super 120B A12B | NVIDIA | 78.9 | 1 | 2026-08-11 |
| 35 | Phi 3.5 Mini Instruct | Microsoft | 68.5 | 1 | 2026-08-23 |
| 36 | Phi 4 Mini Instruct | Microsoft | 67 | 1 | 2026-08-23 |
| 37 | ERNIE 4.5 | Baidu | 51.3 | 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.