llama.cpp on the AMD Radeon AI PRO R9700 32GB
On the AMD Radeon AI PRO R9700 32GB, llama.cpp runs Qwen3.5 35B · 3B active at 4-bit at a median 149.2 tokens per second for generation, from 1 plain run (context not stated).
Speed by model on the AMD Radeon AI PRO R9700 32GB
Plain runs only: one device, one request, no speculative decoding, the whole model in memory. Each row is one model size at one bit class, with the other engines' plain runs of the same model on this card beside it. Context length and build still differ between runs, and each side shows its own.
| Model | llama.cpp | Other engines |
|---|---|---|
| Qwen3.5 35B · 3B active 4-bit | 149.2 t/sone report, context not stated | no other engine reported |
| Qwen3.6 35B · 3B active 5-bit | 84.5 t/sone report, 256K context | no other engine reported |
| Qwen3.8 27B 4-bit | 29.0 t/sone report, context not stated | no other engine reported |
13 llama.cpp reports on the AMD Radeon AI PRO R9700 32GB
Newest first. A run on several cards, with speculative decoding, with part of the model in system RAM or with batched requests says so, since each describes a different setup.
- DeepSeek V4 Flash 284B · 13B active40.0 t/s(8 cards)
- Qwen3.6 27B · Q5_K_M24.9 t/s(2 cards)
- Qwen3.8 125B · 6B active · Q4_K_M36.0 t/s(part in system RAM)
- Qwen3.8 125B · 6B active · IQ3_XXS21.0 t/s(part in system RAM)
- Qwen3.8 27B · Q829.9 t/s(3 cards, speculative decoding)
- Qwen3.5 35B · 3B active · UD-Q4_K_XL149.2 t/s
- DeepSeek V4 Flash 284B · 13B active · UD-IQ4_NL24.1 t/s(2 cards, speculative decoding, part in system RAM)
- Qwen3.8 27B · Q4_K_XL37.0 t/s(2 cards)
- Qwen3.8 125B · 6B active · Q4_K_XL26.6 t/s(part in system RAM)
- Qwen3.8 27B · Q8_K_XL51.6 t/s(2 cards, speculative decoding)
- Qwen3.8 125B · 6B active · Q4_K_M35.0 t/s(2 cards)
- Qwen3.8 27B · Q4_K_XL29.0 t/s
- Qwen3.6 35B · 3B active · Q5_K_M84.5 t/s
Frequently asked
How fast is llama.cpp on the AMD Radeon AI PRO R9700 32GB?
On the AMD Radeon AI PRO R9700 32GB, llama.cpp runs Qwen3.5 35B · 3B active at 4-bit at a median 149.2 tokens per second for generation, from 1 plain run (context not stated).