Compare LLM Models
Select 2-5 models to compare side by side.
Select models to compare
0 / 5 selected
Top downloads
Select 2-5 models
| Attribute |
NousResearch
|
|---|---|
| Company | NousResearch |
| Release Date | 2025-08-25 |
| Parameters (B) | 14 |
| Architecture | Transformer |
| Context Window | 40960 |
| Input Modalities | text |
| Open Source | Yes |
| License | apache-2.0 |
| Score | 0 |
| VRAM (GB) | 29.5 |
| Compute | BF16 inference needs roughly one GPU with ~30GB VRAM (e.g., A100 40GB); 8-bit quantization ~15GB, 4-bit ~8GB on a single consumer GPU |
| Benchmark: AIME-2024 | 55.4 |
| Benchmark: AIME-2025 | 46.8 |
| Benchmark: BBH | 84.4 |
| Benchmark: DROP | 84.5 |
| Benchmark: GPQA-Diamond | 60.2 |
| Benchmark: IFEval | 74.8 |
| Benchmark: LiveCodeBench | 42.5 |
| Benchmark: MATH-500 | 91.1 |
| Benchmark: MMLU | 84.1 |
| Benchmark: MMLU-Pro | 74.3 |
| Benchmark: SimpleQA | 5.5 |
| Pricing (Input) | — |
| Pricing (Output) | — |