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 |
openbmb
|
|---|---|
| Company | openbmb |
| Release Date | 2026-02-11 |
| Parameters (B) | 9 |
| Architecture | Hybrid Sparse-Linear Transformer |
| Context Window | 1000000 |
| Input Modalities | text |
| Open Source | Yes |
| License | apache-2.0 |
| Score | 0 |
| VRAM (GB) | 18 |
| Compute | NVIDIA GPU with ~18GB VRAM for standard BF16 inference; RTX 5090 (32GB) or A6000D recommended for up to 1M-token context |
| Benchmark: AIME-2024 | 83.8 |
| Benchmark: AIME-2025 | 78.3 |
| Benchmark: BBH | 81.5 |
| Benchmark: HumanEval | 95.1 |
| Benchmark: IFEval | 76.3 |
| Benchmark: LiveCodeBench | 52 |
| Benchmark: MBPP | 89.1 |
| Benchmark: MMLU-Pro | 67.0 |
| Pricing (Input) | — |
| Pricing (Output) | — |