Quantized DeepSeek models

1881 GGUF / AWQ / GPTQ versions. These are what you run on your own hardware — full precision DeepSeek V3 needs 671B parameters, a quantized 7B distill fits on a laptop.

ModelBaseMethodDownloadsLikesCommercialLicense
DeepSeek-R1-Distill-Qwen-32B-AWQ MSValdemardiQuantized, Distilled5.1M24✓apache-2.0
DeepSeek-R1-Distill-Qwen-32B-GPTQ-Int4 MStclf90Quantized, Distilled4.7M21✓qwen
DeepSeek-R1-GGUF MSunslothQuantized4.3M146✓mit
deepseek-v4.1-flash-gguf HFDeepSeek-V4.1-Flash Quantized1.2M91?未声明
DeepSeek-R1-Distill-Qwen-7B-GGUF MSunslothQuantized, Distilled686.6K166✓apache-2.0
DeepSeek-R1-Distill-Qwen-32B-QMind-GGUF MSxitaosunQuantized, Distilled617.3K—✓MIT License
DeepSeek-R1-Distill-Qwen-1.5B-GGUF MSunslothQuantized, Distilled543.5K99✓apache-2.0
DeepSeek-R1-Distill-Qwen-32B-GPTQ-Int8 MStclf90Quantized, Distilled476.1K7✓qwen
DeepSeek-R1-Distill-Qwen-7B-Int4-W4A16 量 MSokwindsQuantized, Distilled375.0K7✓Apache License 2.0
DeepSeek-R1-Distill-Qwen-14B-GPTQ-Int4 MStclf90Quantized, Distilled303.5K6✓qwen
DeepSeek-V3-GGUF MSunslothQuantized151.1K29✓mit
DeepSeek-R1-Distill-Qwen-7B-GGUF HFDeepSeek-R1-Distill-Qwen-7Quantized, Distilled141.7K149?未声明
CodeFuse-DeepSeek-33B-4bits MScodefuse-aiQuantized122.9K5✗other
DeepSeek-R1-Distill-Qwen-1.5B-GGUF HFDeepSeek-R1-Distill-Qwen-1Quantized, Distilled117.0K169?未声明
DeepSeek-R1-Distill-Qwen-14B-GGUF MSunslothQuantized, Distilled116.4K95✓apache-2.0
DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF MSISTA-DASLabQuantized86.6K35✓mit
DeepSeek-R1-0528-Qwen3-8B-GGUF MSunslothQuantized85.0K31✓mit
DeepSeek-R1-Distill-Qwen-14B-GGUF HFDeepSeek-R1-Distill-Qwen-1Quantized, Distilled78.4K262?未声明
DeepSeek-R1-Distill-Qwen-32B-GGUF HFDeepSeek-R1-Distill-Qwen-3Quantized, Distilled77.7K333?未声明
DeepSeek-V4-Flash-FP8 MSsgl-projectQuantized77.0K4✓mit
DeepSeek-R1-Distill-Qwen-32B-GGUF MSunslothQuantized, Distilled73.2K107✓apache-2.0
DeepSeek-R1-Distill-Qwen-7B-GGUF HFDeepSeek-R1-Distill-Qwen-7Quantized, Distilled72.5K112?未声明
deepseek-v4.1-flash-gguf HFDeepSeek-V4.1-Flash Quantized68.7K4?未声明
DeepSeek-R1-0528-GGUF MSunslothQuantized68.1K14✓mit
DeepSeek-V4.1-Flash-1017-GGUF MSunslothQuantized62.3K31✓mit
DeepSeek-V3-0324-GGUF MSunslothQuantized62.1K20✓mit
DeepSeek-R1-Distill-Qwen-32B-Int4-W4A16 MSokwindsQuantized, Distilled56.5K18✓Apache License 2.0
DeepSeek-R1-Distill-Qwen-32B-GGUF HFDeepSeek-R1-Distill-Qwen-3Quantized, Distilled54.4K173?未声明
DeepSeek-V4-Flash-0731-GGUF MSunslothQuantized53.3K40✓mit
DeepSeek-R1-Distill-Qwen-14B-GGUF HFDeepSeek-R1-Distill-Qwen-1Quantized, Distilled49.6K147?未声明
DeepSeek-V4-Flash-GGUF MSlovedheartQuantized48.3K5✓MIT License
DeepSeek-R1-Distill-Llama-8B-GGUF MSunslothQuantized, Distilled47.1K46✓llama3.1
Qwen3.5-9B-DeepSeek-V4-Flash-MTP-GGUF MSJackrongQuantized46.8K10✗other
DeepSeek-V4.1-Flash-1017-SF-Int8 MSsiliconflowQuantized46.6K—✓mit
DeepSeek-R1-Distill-Llama-8B-GGUF HFDeepSeek-R1-Distill-Llama-Quantized, Distilled44.9K313?未声明
DeepSeek-R1-awq MScognitivecomputationsQuantized41.1K26✓mit
DeepSeek-R1-GGUF HFDeepSeek-R1 Quantized39.7K1,126?未声明
DeepSeek-R1-Distill-Qwen-1.5B-GGUF HFDeepSeek-R1-Distill-Qwen-1Quantized, Distilled39.7K91?未声明
DeepSeek-V3.2-AWQ MSQuantTrioQuantized37.0K2✓mit
DeepSeek-R1-Distill-Qwen-1.5B-Q4_K_M MSuniontech-yourongQuantized, Distilled36.9K1✓Apache License 2.0
DeepSeek-R1-Distill-Qwen-14B-GGUF HFDeepSeek-R1-Distill-Qwen-1Quantized, Distilled34.5K50?未声明
DeepSeek-V4-Flash-0731-w4a16 MSbaicai1145Quantized33.0K5✗other
DeepSeek-R1-Distill-Llama-70B-GGUF MSunslothQuantized, Distilled32.7K27✓llama3.3
DeepSeek-R1-Distill-Llama-70B-GGUF HFDeepSeek-R1-Distill-Llama-Quantized, Distilled32.1K124?未声明
DeepSeek-V3.1-GGUF MSunslothQuantized26.3K7✓mit
siliconflow/DeepSeek-V4-Flash-0731-SF-IN MSsiliconflowQuantized23.8K—✓mit
DeepSeek-V4.1-Flash-GSQ-RCO-GGUF HFDeepSeek-V4.1-Flash Quantized21.7K22?未声明
deepseek-ai.DeepSeek-V3-0324-GGUF MSDevQuasarQuantized20.2K—?未声明
DeepSeek-V2-Chat-0628-gptq-4bit MSLLM-ResearchQuantized19.8K1?未声明
DeepSeek-V4.1-Flash-GGUF HFDeepSeek-V4.1-Flash Quantized18.8K44?未声明
DeepSeek-R1-Distill-Qwen-7B-GGUF HFDeepSeek-R1-Distill-Qwen-7Quantized, Distilled17.7K90?未声明
DeepSeek-V3-0324-AWQ MScognitivecomputationsQuantized16.8K8✓mit
DeepSeek-R1-Distill-Qwen-7B-GGUF MSlmstudio-communityQuantized, Distilled16.2K28?未声明
DeepSeek-R1-Distill-Llama-8B-GGUF HFDeepSeek-R1-Distill-Llama-Quantized, Distilled16.1K53?未声明
Qwen3.5-9B-DeepSeek-V4-Flash-GGUF MSJackrongQuantized, Distilled15.9K66✓apache-2.0
DeepSeek-R1-Distill-Qwen-32B-GGUF HFDeepSeek-R1-Distill-Qwen-3Quantized, Distilled15.1K39?未声明
DeepSeek-V4-Flash-GGUF MSunslothQuantized14.6K9✓mit
DeepSeek-R1-Distill-Llama-70B-GGUF HFDeepSeek-R1-Distill-Llama-Quantized, Distilled14.5K45?未声明
DeepSeek-V4.1-Flash-ROCMFP23-GGUF HFDeepSeek-V4.1-Flash Quantized14.5K10?未声明
DeepSeek-V2-Lite-IMat-GGUF HFDeepSeek-V2-Lite Quantized14.5K5?未声明
DeepSeek-Coder-V2-Lite-Instruct-GGUF MSAI-ModelScopeQuantized12.7K19✗other
deepseek-ai_DeepSeek-V3.1-GGUF HFDeepSeek-V3.1 Quantized11.9K9?未声明
DeepSeek-R1-0528-Qwen3-8B-AWQ MStclf90Quantized11.5K3✓mit
DeepSeek-V4.1-Flash-GGUF HFDeepSeek-V4.1-Flash Quantized10.9K4?未声明
DeepSeek-R1-Distill-Qwen-14B-GGUF MSlmstudio-communityQuantized, Distilled10.7K23?未声明
DeepSeek-R1-Distill-Qwen-32B-quantized.w HFDeepSeek-R1-Distill-Qwen-3Quantized, Distilled10.6K6?未声明
DeepSeek-V3.1-GGUF HFDeepSeek-V3.1 Quantized10.3K98?未声明
DeepSeek-V3.1-Terminus-gguf-q2ks-mixed-A MSIntelQuantized10.2K—?未声明
DeepSeek-R1-Distill-Qwen-32B-GGUF MSlmstudio-communityQuantized, Distilled9.8K19?未声明
DeepSeek-V3-0324-GGUF-Q4_K_M MSYanfeiSongQuantized9.7K16✓mit
DeepSeek-R1-int4-sym-gguf-q4-0-inc MSOPEAQuantized9.5K2✓mit
DeepSeek-R1-Distill-Qwen-7B-GGUF MSbartowskiQuantized, Distilled9.3K6?未声明
DeepSeek-V3.2-GGUF HFDeepSeek-V3.2 Quantized9.1K41?未声明
DeepSeek-V3.2-NVFP4 MSnv-communityQuantized8.6K2✓mit
DeepSeek-R1-Distill-Llama-70B-GGUF MSlmstudio-communityQuantized, Distilled8.5K21?未声明
FuseO1-DeepSeekR1-QwQ-SkyT1-32B-Preview- MSFuseAIQuantized8.1K—✓apache-2.0
DeepSeek-V4-Pro-Quant-W-INT8-PerChannel- MST-HEADQuantized7.9K1✓mit
DeepSeek-Coder-V2-Instruct-GGUF MSsecond-stateQuantized7.8K2?未声明
DeepSeek-R1-0528-AWQ MScognitivecomputationsQuantized7.6K8✓mit
DeepSeek-R1-GGUF MSlmstudio-communityQuantized7.5K19?未声明

Verified 2026-11-05. Check the file size before downloading — 4-bit and 8-bit differ a lot.

Related

All derived models · Run on devices