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 |
MiniMaxAI
|
|---|---|
| Company | MiniMaxAI |
| Release Date | 2025-01-14 |
| Parameters (B) | 456 |
| Architecture | Hybrid Lightning Attention MoE with ViT |
| Context Window | 1000000 |
| Input Modalities | text, image |
| Open Source | Yes |
| License | |
| Score | 0 |
| VRAM (GB) | 640 |
| Compute | 8x NVIDIA H800 or H20 GPUs (80GB each) with experts_int8 quantization via vLLM |
| Benchmark: MMMU | 68.5 |
| Benchmark: MMMU-Pro | 52.7 |
| Pricing (Input) | — |
| Pricing (Output) | — |