Best local AI models for 10GB VRAM
Ranked by what actually fits at 32K context, computed from real file bytes.
A 10GB card gives you about 9.30 GiB to work with after driver overhead. 12 indexed models fit at 32K context — the largest being Wan2.1-VACE-14B at 17.3B parameters in Q3_K_S.
From the file· fit from summed bytesFrom the file· KV per layer
Fits in 10GB at 32K context
largest quantization that fits, per model
| Model | Modality | Best quant | Params○ | Total◐ | Headroom◐ |
|---|---|---|---|---|---|
| Wan2.2-Animate-14B | video generation | Q3_K | 17.3B | 8.98 GiB | 0.32 GiB |
| Wan2.1-I2V-14B-480P | video generation | Q3_K_M | 16.4B | 8.83 GiB | 0.47 GiB |
| Bernini-R | video generation | Q4_K_S | 14.3B | 8.99 GiB | 0.31 GiB |
| Wan2.2-Distill-Models | video generation | Q4_K_S | 14.3B | 8.99 GiB | 0.31 GiB |
| Wan2.2-TI2V-5B | video generation | Q8_0 | 5.0B | 5.87 GiB | 3.43 GiB |
| Wan2.1-T2V-14B | video generation | Q4_0 | 14.3B | 9.25 GiB | 0.05 GiB |
| Wan2.1-I2V-14B-720P | video generation | Q3_K_M | 16.4B | 8.83 GiB | 0.47 GiB |
| Wan2.1-VACE-14B | video generation | Q3_K_S | 17.3B | 8.14 GiB | 1.16 GiB |
| Wan2.1-FLF2V-14B-720P | video generation | Q3_K_M | 16.4B | 8.83 GiB | 0.47 GiB |
| HunyuanVideo-1.5 | video generation | Q8_0 | 8.3B | 9.22 GiB | 0.08 GiB |
| Wan2.2-TI2V-5B-Turbo | video generation | Q8_0 | 5.0B | 5.87 GiB | 3.43 GiB |
| SkyReels-V2-DF-14B-540P | video generation | Q4_K_S | 14.3B | 8.99 GiB | 0.31 GiB |
This page models a generic 10GB accelerator, so it answers what fits rather than how fast it runs. For tokens per second you need a specific card — pick one from hardware, where bandwidth is known.