DeepSeek V4 FLASH

No machine on this network has measured this model yet. Everything below is what the catalog knows exists — where the weights come from, how large each build is, and whose work it is. Nothing here is a claim about what the model can do; that only appears once a machine has run it and signed the result.

hf.co/antirez/deepseek-v4-gguf:F32ollama · F32 · 3.5 GiB · quantizer (antirez)

unsloth/DeepSeek-V4-Flash-GGUF:IQ1_Sllamacpp · IQ1_S · 76.9 GiB · quantizer (unsloth)

unsloth/DeepSeek-V4-Flash-GGUF:IQ2_XXSllamacpp · IQ2_XXS · 84.6 GiB · quantizer (unsloth)

unsloth/DeepSeek-V4-Flash-GGUF:IQ3_XXSllamacpp · IQ3_XXS · 95.9 GiB · quantizer (unsloth)

mlx-community/DeepSeek-V4-Flashmlx · 8BIT · 144.4 GiB · quantizer (mlx-community)

unsloth/DeepSeek-V4-Flash-GGUF:Q4_K_XLllamacpp · Q4_K_XL · 144.4 GiB · quantizer (unsloth)

sgl-project/DeepSeek-V4-Flash-FP8vllm · FP8 · 273.8 GiB · quantizer (sgl-project)

derivatives — somebody else's change to these weights

Fine-tunes, merges and abliterations of this member, each its own model with its own page. None of their numbers are counted above.

Huihui-DeepSeek-V4-Flash-abliterated-ds4-GGUFnot measured DeepSeek-V4-Flash-4bitnot measured DeepSeek-V4-Flash-VQ-3.2bpwnot measured