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 |
OpenGVLab
|
|---|---|
| Company | OpenGVLab |
| Release Date | 2025-08-26 |
| Parameters (B) | 30.9 |
| Architecture | ViT-MLP-MoE Transformer |
| Context Window | 32768 |
| Input Modalities | text, image |
| Open Source | Yes |
| License | apache-2.0 |
| Score | 0 |
| VRAM (GB) | 80 |
| Compute | Single NVIDIA A100 80GB GPU (bf16); supports 8-bit quantization and multi-GPU deployment via vLLM/LMDeploy |
| Benchmark: AIME-2024 | 79.4 |
| Benchmark: AIME-2025 | 62.7 |
| Benchmark: BBH | 83.8 |
| Benchmark: C-Eval | 83.2 |
| Benchmark: GPQA | 68.2 |
| Benchmark: GSM8K | 91.4 |
| Benchmark: HellaSwag | 96.3 |
| Benchmark: HumanEval | 96.3 |
| Benchmark: IFEval | 74.3 |
| Benchmark: MATH | 93.6 |
| Benchmark: MATH-500 | 96.6 |
| Benchmark: MMLU | 83 |
| Benchmark: MMLU-Pro | 75.3 |
| Benchmark: MMMU | 75.6 |
| Pricing (Input) | — |
| Pricing (Output) | — |