What can the Radeon AI PRO R9700 run?
| Text models: 233 | Visual models: 15 | Strongest first |
- Memory
- 32 GB
- Bandwidth
- 640 GB/s
- Maker
- AMD
Open the calculator on this machineEvery model for 32 GB of VRAM
Text models that run on it
| Model | Best fit | Context | Speed | Strongest at |
|---|---|---|---|---|
| Qwen3.8 27B | 8-bit | 24K | 13 tok/s | Chat, Coding, Agents, Docs |
| Muse Glimmer 30B | 6-bit | 128K | 16 tok/s | Docs |
| Qwen3.6 35B A3B | 6-bit | 96K | 98 tok/s | Chat, Coding, Agents, Docs |
| Gemma 4 31B | 6-bit | 112K | 14 tok/s | Chat, Coding, Agents, Docs |
| Gemma 4 26B A4B | 8-bit | 256K | 76 tok/s | Chat, Coding, Agents, Docs |
| Qwen3.5 9B | 8-bit | 256K | 39 tok/s | Chat, Coding, Agents, Docs |
| GLM-4.7 Flash | 6-bit | 104K | 94 tok/s | Chat, Coding, Agents |
| Gemma 4 12B | 8-bit | 256K | 28 tok/s | Chat, Coding, Docs |
| Nemotron 3.5 Lightning 30B A3B | 6-bit | 256K | 89 tok/s | Agents, Docs |
| Devstral Small 2 24B | FP8 | 32K | 15 tok/s | Coding |
| LFM2.5 24B A2B | 8-bit | 128K | 109 tok/s | Agents |
| Mistral Small 3.2 24B | 8-bit | 36K | 15 tok/s | Chat, Coding, Agents |
| Ministral 3 14B | 8-bit | 96K | 26 tok/s | Coding, Agents, Docs |
| Gemma 4 E4B | 8-bit | 128K | 61 tok/s | Chat, Coding, Docs |
| gpt-oss 20B | MXFP4 | 128K | 113 tok/s | Agents |
| Ministral 3 8B | 8-bit | 152K | 40 tok/s | Docs |
| Olmo 3.1 32B | 6-bit | 64K | 14 tok/s | Chat, Coding, Agents, Docs |
| Llama 3.1 8B | 8-bit | 128K | 45 tok/s | Chat, Coding, Agents, Docs |
| Qwen3 14B | 8-bit | 40K | 25 tok/s | Not rated |
| Qwen3 VL 8B Instruct | 8-bit | 144K | 41 tok/s | Text and images |
| Qwen3 8B | 8-bit | 40K | 44 tok/s | Not rated |
| Qwen2.5 7B Instruct | 8-bit | 32K | 48 tok/s | Not rated |
| Qwen2.5 VL 7B Instruct | 8-bit | 125K | 44 tok/s | Text and images |
| Qwen3 VL 4B Instruct | 8-bit | 176K | 71 tok/s | Text and images |
| Qwen3 4B | 8-bit | 40K | 77 tok/s | Not rated |
| Qwen3 32B | 6-bit | 16K | 14 tok/s | Not rated |
| Qwen3 4B Instruct 2507 | 8-bit | 176K | 77 tok/s | Not rated |
| Qwen3.6 27B | 8-bit | 24K | 13 tok/s | Text and images |
| Qwen2.5 Coder 14B Instruct | 8-bit | 32K | 24 tok/s | Coding |
| Qwen3 Coder 30B A3B Instruct | 6-bit | 64K | 93 tok/s | Coding |
| Qwen3.5 27B | 8-bit | 24K | 13 tok/s | Text and images |
| Qwen3.5 35B A3B | 6-bit | 96K | 98 tok/s | Text and images |
| NVIDIA Nemotron 3 Nano 30B A3B BF16 | 6-bit | 256K | 89 tok/s | Not rated |
| Qwen2.5 Coder 7B Instruct | 8-bit | 32K | 48 tok/s | Coding |
| Mistral 7B Instruct V0.2 | 8-bit | 32K | 47 tok/s | Not rated |
| Qwen3 30B A3B | 6-bit | 40K | 93 tok/s | Not rated |
| Qwen3 30B A3B Instruct 2507 | 6-bit | 64K | 93 tok/s | Not rated |
| Laguna XS.2 | 6-bit | 104K | 100 tok/s | Not rated |
| OTel 2.0 LLM 31B IT | 6-bit | 112K | 14 tok/s | Not rated |
| Qwen2.5 32B Instruct | 6-bit | 16K | 14 tok/s | Not rated |
| MiniCPM V 4.7 35B A3B | 6-bit | 120K | 98 tok/s | Text and images |
| Qwen2.5 14B Instruct | 8-bit | 32K | 24 tok/s | Not rated |
| Laguna XS 2.1 | 6-bit | 104K | 100 tok/s | Not rated |
| Diffusiongemma 26B A4B IT | 8-bit | 184K | 73 tok/s | Text and images |
| Qwen2.5 VL 32B Instruct | 5-bit | 28K | 16 tok/s | Text and images |
| Qwen2.5 Coder 32B Instruct | 6-bit | 16K | 14 tok/s | Coding |
| Qwen2 VL 7B Instruct | 8-bit | 32K | 44 tok/s | Text and images |
| Ornith 1.5 9B | 8-bit | 256K | 39 tok/s | Not rated |
| Qwen3 VL 30B A3B Instruct | 6-bit | 60K | 93 tok/s | Text and images |
| Mistral 7B V0.1 | 8-bit | 32K | 47 tok/s | Not rated |
| Meta Llama 3 8B Instruct Abliterated V3 | 8-bit | 8K | 45 tok/s | Not rated |
| EXAONE 3.5 7.8B Instruct | 8-bit | 32K | 45 tok/s | Not rated |
| Saiga Llama3 8B | 8-bit | 8K | 45 tok/s | Not rated |
| DeepSeek R1 0528 Qwen3 8B | 8-bit | 128K | 44 tok/s | Not rated |
| Gemma 3 12B IT | 8-bit | 128K | 30 tok/s | Text and images |
| Phi 3.5 Mini Instruct | 8-bit | 64K | 84 tok/s | Not rated |
| DeepSeek R1 Distill Qwen 32B | 6-bit | 16K | 14 tok/s | Not rated |
| Qwen3 4B Thinking 2507 | 8-bit | 176K | 77 tok/s | Not rated |
| Gemma 3 27B IT | 8-bit | 20K | 13 tok/s | Text and images |
| Phi 4 Mini Instruct | 8-bit | 128K | 82 tok/s | Not rated |
Best fit is the highest quality, up to 8-bit, that leaves the use case’s context. Ranked by our use-case scores, then by Hugging Face downloads. Speeds are rough, plus or minus 25%. The top 60 of 233 are shown.
Image and video models
| Model | Best fit | Per output | Strongest at |
|---|---|---|---|
| MiniMax H3 | 8-bit | 8.9 min | Video |
| LTX-2.5 | 8-bit | 71 s | Video |
| FLUX.1 [dev] | 16-bit | 17 s | Images |
| Qwen-Image | 8-bit | 5.1 min | Images |
| FLUX.2 [dev] | 6-bit | 2.6 min | Images |
| Qwen-Image 2.1 | 16-bit | 2.2 min | Images |
| Wan2.2 I2V A14B | 8-bit | 46 min | Video |
| FLUX.2 [klein] 9B | 16-bit | 4.7 s | Images |
| Wan2.1 I2V 14B | 8-bit | 46 min | Video |
| Wan 2.1 T2V-14B | 8-bit | 57 min | Video |
| Krea 2 Turbo | 16-bit | 17 s | Images |
| Stable Diffusion 3.5 Large | 16-bit | 70 s | Images |
| Qwen Image 2512 | 8-bit | 5.1 min | Images |
| Wan 2.2 A14B | 8-bit | 46 min | Video |
| HunyuanVideo 1.5 | 16-bit | 56 min | Video |
Only just too big
- Nemotron 3.5 Bolt 74B A9B
- Qwen3 Coder Next 80B
- Llama 3.3 70B
- Qwen2.5 72B Instruct
- Qwen3 Next 80B A3B Instruct
- Qwen2.5 VL 72B Instruct
- Solar Open 100B
- Qwen3 Next 80B A3B Thinking
- DeepSeek R1 Distill Llama 70B
- Qwen2 72B Instruct
- Qwen2 VL 72B Instruct
- Llama 3.1 70B
- GLM 4.5V
- InternVL3 78B
These need up to a quarter more memory than the Radeon AI PRO R9700 has. Spilling into system RAM or a lower precision can get them running, slowly.
Questions
What is the best LLM for the Radeon AI PRO R9700?
Of the open models we track, the strongest that run well on the Radeon AI PRO R9700 are Qwen3.8 27B, Muse Glimmer 30B, Qwen3.6 35B A3B. 233 text models fit it in all.
Calculated from published specs, not benchmarked. All machines