DeepSeek-V4-Flash-Vision-Exp

8 records across 2 hardware.

hardwaremodelquantbackendtok/spp tok/sWttft sctx/tgdateclassflagssource
M5 UltraDeepSeek-V4-Flash-Vision-Exp 4bitoMLX38.0621.1––10242026-10-09community–source
M5 UltraDeepSeek-V4-Flash-Vision-Exp 4bitoMLX33.9946.5––40962026-10-09community–source
RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GBDeepSeek-V4-Flash-Vision-Exp 284B-A13BUD-IQ3_XXSllama.cpp (CUDA)10.3249.5––2620002026-10-05community–source
RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GBDeepSeek-V4-Flash-Vision-Exp 284B-A13BUD-IQ3_XXSllama.cpp (CUDA)9.9045.7––2620002026-10-05community–source
RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GBDeepSeek-V4-Flash-Vision-Exp 284B-A13BUD-IQ3_XXSllama.cpp (CUDA)9.7144.3––2620002026-10-05community–source
RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GBDeepSeek-V4-Flash-Vision-Exp 284B-A13BUD-IQ3_XXSllama.cpp (CUDA)9.0944.2––2620002026-10-05community–source
RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GBDeepSeek-V4-Flash-Vision-Exp 284B-A13BUD-IQ3_XXSllama.cpp (CUDA)6.1045.1––2620002026-10-05community–source
RTX 5070 Ti 16 GB + RTX 4060 Ti 16 GBDeepSeek-V4-Flash-Vision-Exp 284B-A13BUD-IQ3_XXSllama.cpp (CUDA)5.3643.7––2620002026-10-05community–source

← all models