§01·spec · /gpus
RTX 3090 Ti
nvidia30 series24GB VRAM
41 open-weights AI models run on the RTX 3090 Ti — see which fit, how fast they go, and the VRAM each needs.
§02·models that run on this GPU
41 totalLLM · 14
| Model | Best speed | Min VRAM | Works | Evidence | |
|---|---|---|---|---|---|
| Qwen3 30B-A3B | 166.9tokens/s | 24GB | ✓ | 2benchesrecipe | check ↗ |
| gpt-oss 20B | 160.3tokens/s | 16GB | ✓ | 2benchesrecipe | check ↗ |
| Qwen3-8B | 123.7tokens/s | 6GB | ✓ | 2benchesrecipe | check ↗ |
| Qwen3 14B | 76.2tokens/s | 10GB | ✓ | 2benchesrecipe | check ↗ |
| Qwen3 32B | 38tokens/s | 22GB | ✓ | 2benchesrecipe | check ↗ |
| Apodex 1.1 mini | 24GB | ✓ | recipe | check ↗ | |
| DeepSeek R1 Distill 14B | 12GB | ✓ | recipe | check ↗ | |
| Devstral Small 2 (24B) | 24GB | ✓ | recipe | check ↗ | |
| Laguna XS 2.1 | 24GB | ✓ | recipe | check ↗ | |
| Llama 3.1 8B | 10GB | ✓ | recipe | check ↗ | |
| Mistral Small 3.2 24B | 24GB | ✓ | recipe | check ↗ | |
| Nanbeige4.2 3B | 24GB | ✓ | recipe | check ↗ | |
| North Mini Code 1.0 | 24GB | ✓ | recipe | check ↗ | |
| Ornith 1.0 35B | 24GB | ✓ | recipe | check ↗ |
- Best speed
- 166.9tokens/s
- Min VRAM
- 24GB
- Evidence
- 2benches
- Best speed
- 160.3tokens/s
- Min VRAM
- 16GB
- Evidence
- 2benches
- Best speed
- 123.7tokens/s
- Min VRAM
- 6GB
- Evidence
- 2benches
- Best speed
- 76.2tokens/s
- Min VRAM
- 10GB
- Evidence
- 2benches
- Best speed
- 38tokens/s
- Min VRAM
- 22GB
- Evidence
- 2benches
- Best speed
- Min VRAM
- 24GB
- Best speed
- Min VRAM
- 12GB
- Best speed
- Min VRAM
- 24GB
- Best speed
- Min VRAM
- 24GB
- Best speed
- Min VRAM
- 10GB
- Best speed
- Min VRAM
- 24GB
- Best speed
- Min VRAM
- 24GB
- Best speed
- Min VRAM
- 24GB
- Best speed
- Min VRAM
- 24GB
Multimodal · 7
| Model | Best speed | Min VRAM | Works | Evidence | |
|---|---|---|---|---|---|
| Fara1.5-27B | 24GB | ✓ | recipe | check ↗ | |
| Fara1.5-4B | 16GB | ✓ | recipe | check ↗ | |
| Fara1.5-9B | 24GB | ✓ | recipe | check ↗ | |
| Gemma 4 E4B-IT | 20GB | ✓ | recipe | check ↗ | |
| Gemma 4 26B MoE | 18GB | ✓ | recipe | check ↗ | |
| Muse Glimmer 30B | 24GB | ✓ | recipe | check ↗ | |
| Qwen3.8 27B | 24GB | ✓ | recipe | check ↗ |
Image · 10
| Model | Best speed | Min VRAM | Works | Evidence | |
|---|---|---|---|---|---|
| Flux.1 Dev | 1.27it/s | 24GB | ✓ | 2benchesrecipe | check ↗ |
| Chroma V48 | 24GB | ✓ | recipe | check ↗ | |
| Flux.2-Klein-4B | 13GB | ✓ | recipe | check ↗ | |
| HiDream-O1-Image | 20GB | ✓ | recipe | check ↗ | |
| Juggernaut Z | 16GB | ✓ | recipe | check ↗ | |
| Krea 2 | 16GB | ✓ | recipe | check ↗ | |
| LongCat Image | 18GB | ✓ | recipe | check ↗ | |
| Qwen-Image | 13GB | ✓ | recipe | check ↗ | |
| Qwen-Image-2.1 | 12GB | ✓ | recipe | check ↗ | |
| Z-Image Turbo | 16GB | ✓ | recipe | check ↗ |
- Best speed
- 1.27it/s
- Min VRAM
- 24GB
- Evidence
- 2benches
- Best speed
- Min VRAM
- 24GB
- Best speed
- Min VRAM
- 13GB
- Best speed
- Min VRAM
- 20GB
- Best speed
- Min VRAM
- 16GB
- Best speed
- Min VRAM
- 16GB
- Best speed
- Min VRAM
- 18GB
- Best speed
- Min VRAM
- 13GB
- Best speed
- Min VRAM
- 12GB
- Best speed
- Min VRAM
- 16GB
Video · 8
| Model | Best speed | Min VRAM | Works | Evidence | |
|---|---|---|---|---|---|
| CogVideoX 1.5 | 14GB | ✓ | recipe | check ↗ | |
| HunyuanVideo 1.5 | 24GB | ✓ | recipe | check ↗ | |
| LightX2V | 14GB | ✓ | recipe | check ↗ | |
| LTX-2.5 | 24GB | ✓ | recipe | check ↗ | |
| MiniMax H3 (Hailuo 3) | 12GB | ✓ | recipe | check ↗ | |
| Mochi 1 | 22GB | ✓ | recipe | check ↗ | |
| Wan 2.2 TI2V-5B | 24GB | ✓ | recipe | check ↗ | |
| Wan 2.2 14B | 24GB | ✓ | recipe | check ↗ |
3D · 1
| Model | Best speed | Min VRAM | Works | Evidence | |
|---|---|---|---|---|---|
| Waypoint 1.5 | 12GB | ✓ | recipe | check ↗ |
§03·tested recipes
showing 6 of 41- imageintermediate12GB+recipe
Qwen-Image-2.1 on RTX 3090 Ti: int8 ComfyUI setup, what stays resident in 24 GB, 2K and edits
- musicintermediate16GB+recipe
YuE2-3B on RTX 3090 Ti: full songs, and the host RAM it needs
- llmintermediate24GB+recipe
Apodex 1.1 mini on RTX 3090 Ti: 128K Agent Context in Q4_K_M, and the CUDA-Graph Caveat
- multimodalintermediate24GB+recipe
Muse Glimmer 30B on RTX 3090 Ti: vision + DFlash speculation at full 131K context
- multimodalintermediate24GB+recipe
Qwen3.8-27B on RTX 3090 Ti: 128K-context vision chat via llama.cpp Q4_K_M
- videoadvanced24GB+recipe
LTX-2.5 on RTX 3090 Ti: 22B audio-video on Ampere via the official ComfyUI INT8 checkpoint