DeepSeek-V4-Flash-Vision-Exp
8 records across 2 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M5 Ultra | DeepSeek-V4-Flash-Vision-Exp | 4bit | oMLX | 38.0 | 621.1 | – | – | 1024 | 2026-10-09 | community | – | source |
| M5 Ultra | DeepSeek-V4-Flash-Vision-Exp | 4bit | oMLX | 33.9 | 946.5 | – | – | 4096 | 2026-10-09 | community | – | source |
| RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GB | DeepSeek-V4-Flash-Vision-Exp 284B-A13B | UD-IQ3_XXS | llama.cpp (CUDA) | 10.32 | 49.5 | – | – | 262000 | 2026-10-05 | community | – | source |
| RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GB | DeepSeek-V4-Flash-Vision-Exp 284B-A13B | UD-IQ3_XXS | llama.cpp (CUDA) | 9.90 | 45.7 | – | – | 262000 | 2026-10-05 | community | – | source |
| RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GB | DeepSeek-V4-Flash-Vision-Exp 284B-A13B | UD-IQ3_XXS | llama.cpp (CUDA) | 9.71 | 44.3 | – | – | 262000 | 2026-10-05 | community | – | source |
| RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GB | DeepSeek-V4-Flash-Vision-Exp 284B-A13B | UD-IQ3_XXS | llama.cpp (CUDA) | 9.09 | 44.2 | – | – | 262000 | 2026-10-05 | community | – | source |
| RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GB | DeepSeek-V4-Flash-Vision-Exp 284B-A13B | UD-IQ3_XXS | llama.cpp (CUDA) | 6.10 | 45.1 | – | – | 262000 | 2026-10-05 | community | – | source |
| RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GB | DeepSeek-V4-Flash-Vision-Exp 284B-A13B | UD-IQ3_XXS | llama.cpp (CUDA) | 5.36 | 43.7 | – | – | 262000 | 2026-10-05 | community | – | source |