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 |
ibm-granite
|
|---|---|
| Company | ibm-granite |
| Release Date | 2025-10-28 |
| Parameters (B) | 0.3 |
| Architecture | Hybrid Mamba-2/Transformer |
| Context Window | 32000 |
| Input Modalities | text |
| Open Source | Yes |
| License | apache-2.0 |
| Score | 5 |
| VRAM (GB) | 0.8 |
| Compute | Runs on consumer CPUs with 8GB+ RAM or any GPU with ~1GB VRAM; compatible with vLLM, Ollama, llama.cpp, MLX, and Docker |
| Benchmark: AGIEval | 29.6 |
| Benchmark: BBH | 33.3 |
| Benchmark: GPQA | 26.1 |
| Benchmark: GSM8K | 39.3 |
| Benchmark: HumanEval | 38 |
| Benchmark: IFEval | 61.6 |
| Benchmark: MBPP | 49 |
| Benchmark: MMLU | 36.2 |
| Benchmark: MMLU-Pro | 14.4 |
| Pricing (Input) | — |
| Pricing (Output) | — |