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 | 2026-05-11 |
| Parameters (B) | 1.3 |
| Architecture | Multimodal Transformer |
| Context Window | 262144 |
| Input Modalities | text, image, video |
| Open Source | Yes |
| License | apache-2.0 |
| Score | 6 |
| VRAM (GB) | 3 |
| Compute | GPTQ W4A16 quantized weights require ~3 GB GPU VRAM; also deployable on consumer GPUs via vLLM/SGLang, mobile devices (iOS/Android/HarmonyOS), or CPU with GGUF (~2 GB) |
| Benchmark: GPQA-Diamond | 30.5 |
| Benchmark: HLE | 4.9 |
| Benchmark: IFEval | 26.7 |
| Benchmark: LiveCodeBench | 6.3 |
| Benchmark: MMMU-Pro | 38 |
| Benchmark: Terminal-Bench-2.1 | 0 |
| Pricing (Input) | — |
| Pricing (Output) | — |