111
Total multimodal profiles
Multimodal Models Hub
Use this hub to shortlist image-capable models by memory budget. Start with local-first options, then escalate to larger cloud-first profiles only when your tasks need heavier reasoning depth.
14
Local-first picks
10
Cloud-first picks
8
Size classes covered
Local-first multimodal picks (<=24GB optimal)
| Model | VRAM min/optimal | 3090 tok/s | Detail |
|---|---|---|---|
| LLaVA 7B Q4 | 8GB / 10GB | 30 | Open |
| Gemma 3 270M Q4 | 2GB / 10GB | 48 | Open |
| LLaVA 7B Q5 | 10GB / 12GB | 27 | Open |
| Qwen3 VL 2B Q4 | 2GB / 12GB | 42 | Open |
| Qwen3 VL 2B CLOUD | 4GB / 12GB | 42 | Open |
| Gemma 3n E2B Q4 | 2GB / 12GB | 42 | Open |
| Gemma 3 1B Q4 | 2GB / 12GB | 48 | Open |
| Gemma 3 270M Q5 | 2GB / 12GB | 43.2 | Open |
| Llama 3.2 Vision 11B Q4 | 12GB / 14GB | 21 | Open |
| Gemma 3 4B Q4 | 4GB / 14GB | 36 | Open |
| Qwen3 VL 4B Q4 | 4GB / 14GB | 36 | Open |
| Qwen3 VL 4B CLOUD | 6GB / 14GB | 36 | Open |
| Gemma 3n E4B Q4 | 4GB / 14GB | 36 | Open |
| Qwen2.5 VL 3B Q4 | 4GB / 14GB | 36 | Open |
Cloud-first multimodal picks (>24GB optimal)
| Model | VRAM min/optimal | Cloud fallback | Detail |
|---|---|---|---|
| Llama 4 128X17B Q4 | 418GB / 428GB | H100/H200 class | Open |
| Llama 4 128X17B Q5 | 420GB / 430GB | H100/H200 class | Open |
| Llama 4 128X17B Q8 | 424GB / 434GB | H100/H200 class | Open |
| Llama 4 128X17B FP16 | 430GB / 442GB | H100/H200 class | Open |
| Qwen3.5 397B-A17B CLOUD | 215GB / 223GB | H100/H200 class | Open |
| Llama 4 16X17B Q4 | 213GB / 223GB | H100/H200 class | Open |
| Llama 4 16X17B Q5 | 215GB / 225GB | H100/H200 class | Open |
| Llama 4 16X17B Q8 | 219GB / 229GB | H100/H200 class | Open |
| Llama 4 16X17B FP16 | 225GB / 237GB | H100/H200 class | Open |
| Qwen3 VL 235B Q4 | 138GB / 148GB | H100/H200 class | Open |