LFM3-VL-2B
Side-by-side specs, memory needs, licence, runtimes and benchmarks. Add up to 4 models.
--vram
--gpus
--quant
--add
On your GPU
- LFM3-VL-2B
| Property | |
|---|---|
| Overview | |
| Task | Vision |
| Parameters | 1.9B |
| Active parameters | dense |
| Context length | 125K |
| Input → output | |
| Released | 19 Oct 2026 |
| Licence | lfm1.0 |
| Gated download | no |
| Memory and speed | |
| Fits your hardwarepick a GPU above | ≈2.0 GB |
| VRAM at Q4 | ≈2.0 GB |
| VRAM at Q8 | ≈2.9 GB |
| VRAM at FP16 | ≈4.6 GB |
| Est. generation speedpick a GPU above | — |
| Popularity | |
| Downloads / 30 days | 30.7K |
| Likes | 288 |
| Runs with | |
| Jan | |
| KoboldCpp | |
| llama.cpp | |
| LM Studio | |
| LocalAI | |
| MLX | |
| Ollama | |
| ONNX Runtime | |
| Transformers | |
| Ollama tag | lfm3-vl:2b |
| Architecture | |
| Architecture | Lfm3VlForConditionalGeneration |
| Layers | 24 |
| Attention / KV heads | 16 / 4 |
| Vocabulary | 128 000 |
| Weights dtype | bfloat16 |
| GGUF available | yes |
Keep comparing
Memory figures are estimates for an 8K context at the chosen quantisation (Q4_K_M for LLMs by default). Speeds are rough, memory-bound estimates. Self-reported benchmark scores come from the model authors.