DeepSeek V3.2 BASE

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.

unsloth/DeepSeek-V3.2-GGUF:IQ1_Sllamacpp · IQ1_S · 171.5 GiB · quantizer (unsloth)

unsloth/DeepSeek-V3.2-GGUF:IQ2_XXSllamacpp · IQ2_XXS · 202.2 GiB · quantizer (unsloth)

unsloth/DeepSeek-V3.2-GGUF:IQ3_XXSllamacpp · IQ3_XXS · 254.4 GiB · quantizer (unsloth)

QuantTrio/DeepSeek-V3.2-AWQvllm · AWQ · 337.1 GiB · quantizer (QuantTrio)

unsloth/DeepSeek-V3.2-GGUF:Q4_0llamacpp · Q4_0 · 354.0 GiB · quantizer (unsloth)

amd/DeepSeek-V3.2-mxfp4vllm · QUARK · 377.0 GiB · quantizer (amd)

unsloth/DeepSeek-V3.2-GGUF:Q4_K_Mllamacpp · Q4_K_M · 377.6 GiB · quantizer (unsloth)

mlx-community/DeepSeek-V3.2-mlx-5bitmlx · 5BIT · 430.3 GiB · quantizer (mlx-community)

deepseek-ai/DeepSeek-V3.2vllm · FP8 · 642.1 GiB · vendor (deepseek-ai)

unsloth/DeepSeek-V3.2-GGUF:Q8_0llamacpp · Q8_0 · 664.3 GiB · quantizer (unsloth)

unsloth/DeepSeek-V3.2-GGUF:BF16llamacpp · BF16 · 1250.1 GiB · quantizer (unsloth)

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.

DeepSeek-V3.2-RAM-350GB-MLXnot measured DeepSeek-V3.2-REAP-345B-A37Bnot measured DeepSeek-V3.2_bf16not measured DeepSeek-V3.2-4bitnot measured DeepSeek-V3.2-NVFP4not measured DeepSeek-V3.2-MLX-4bitnot measured