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 |
tiiuae
|
|---|---|
| Company | tiiuae |
| Release Date | 2024-12-17 |
| Parameters (B) | 7 |
| Architecture | Transformer |
| Context Window | 32768 |
| Input Modalities | text |
| Open Source | Yes |
| License | other |
| Score | 0 |
| VRAM (GB) | 14 |
| Compute | BF16 inference needs ~14GB VRAM on a single GPU (16GB recommended with KV cache); Q4 quantization runs on 8GB cards. Pretrained on 1,024 NVIDIA H100 GPUs. |
| Benchmark: ARC | 62.6 |
| Benchmark: BBH | 52.4 |
| Benchmark: GPQA | 31.9 |
| Benchmark: GSM8K | 81.4 |
| Benchmark: IFEval | 76.5 |
| Benchmark: MATH | 29.4 |
| Benchmark: MMLU | 70.5 |
| Benchmark: MMLU-Pro | 40.7 |
| Pricing (Input) | — |
| Pricing (Output) | — |