MiniMax M2.5 MiniMax-M2.5 / b8118 Vulkan / PP512 TG128 at prepared depth8000
AMD Radeon 8060S gfx1151 (Strix Halo; CPU SKU and installed RAM unknown) · llama.cpp
- generation:
- 25.9 tokens/s
- prompt processing (prefill):
- 119.4 tokens/s
- quant:
- UD-IQ3_XXS (GGUF)
- kv:
- Q8
- flash attention:
- on
- mtp (multi-token prediction):
- off
Reported by the source; GPU count, offloading and concurrent requests can change these figures. Prompt-processing speed is input throughput, not time to first token. Check the full setup before comparing.
User reports MiniMax-M2.5 UD-IQ3_XXS at 25.9 t/s generation and 119.4 t/s prompt processing on a single Radeon 8060S gfx1151 device. Setup is llama.cpp build b8118 with q8_0 KV cache, -ngl 999, -fa 1, -ub 2048 -b 2048, and depth 8000. The headline figures come from the Vulkan/RADV arm, where PP512 measured 119.44 ± 4.23 and TG128 25.87 ± 0.52 tok/s; the ROCm/HIP arm reached 61.32 and 16.58 tok/s. Depth 8000 is prepared before timing and excluded from the numerator.