Best AI models for every VRAM size
Pick how much GPU memory you have. Each page ranks the strongest open-weight models that fit fully in it (chat, code, reasoning, vision, image generation, speech and embeddings), with the best quantisation and an estimated speed.
- 4GBTop chat pickgranite-4.2-3b
- 6GBTop chat pickgranite-4.2-3b
- 8GBTop chat pickQwen3-8B
- 10GBTop chat pickQwen3-8B
- 12GBTop chat pickQwen3-8B
- 16GBTop chat pickgpt-oss-20b
- 20GBTop chat pickLFM2.5-24B-A2B
- 24GBTop chat pickOrnith-1.5-35B-A3B
- 32GBTop chat pickOrnith-1.5-35B-A3B
- 48GBTop chat pickOrnith-1.5-35B-A3B
- 64GBTop chat pickOrnith-1.5-35B-A3B
- 80GBTop chat pickGLM-4.5-Air
- 96GBTop chat pickLaguna-S-2.1
- 128GBTop chat pickLaguna-S-2.1
Memory figures include an 8K-token KV cache and runtime overhead. Speeds assume a typical card of each size; your GPU’s own page has estimates for its exact bandwidth.