HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks
Home/Orgs/deepseek-ai

deepseek-ai

publisher · 101 variants · 94 roots

huggingface
https://huggingface.co/deepseek-ai

Root models (94)

rootfamilyvariantsbenchmarks
deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5Bqwen1410
deepseek-ai/DeepSeek-R1-0528-Qwen3-8Bqwen122
deepseek-ai/DeepSeek-R1-Distill-Qwen-7Bqwen127
deepseek-ai/DeepSeek-R1-Distill-Qwen-32Bqwen116
deepseek-ai/DeepSeek-R1deepseek115
deepseek-ai/DeepSeek-R1-Distill-Qwen-14Bqwen108
deepseek-ai/DeepSeek-R1-Distill-Llama-70Bllama95
deepseek-ai/DeepSeek-R1-Distill-Llama-8Bllama85
deepseek-ai/DeepSeek-R1-0528deepseek72
deepseek-ai/deepseek-coder-6.7b-basedeepseek60
deepseek-ai/DeepSeek-V4-Flashdeepseek623
deepseek-ai/DeepSeek-V3-0324deepseek68
deepseek-ai/deepseek-coder-6.7b-instructdeepseek50
deepseek-ai/deepseek-coder-1.3b-instructdeepseek40
deepseek-ai/deepseek-coder-33b-instructdeepseek40
deepseek-ai/DeepSeek-V3.2-Exp-Basedeepseek48
deepseek-ai/DeepSeek-V3.1-Basedeepseek47
deepseek-ai/DeepSeek-V4-Prodeepseek423
deepseek-ai/DeepSeek-OCRdeepseek30
deepseek-ai/DeepSeek-Coder-V2-Lite-Instructdeepseek30
deepseek-ai/deepseek-coder-7b-instruct-v1.5deepseek30
deepseek-ai/deepseek-llm-7b-chatdeepseek30
deepseek-ai/deepseek-llm-7b-basedeepseek30
deepseek-ai/deepseek-coder-1.3b-basedeepseek30
deepseek-ai/DeepSeek-V2deepseek311
deepseek-ai/DeepSeek-Coder-V2-Lite-Basedeepseek30
deepseek-ai/DeepSeek-V2.5deepseek35
deepseek-ai/DeepSeek-Coder-V2-Basedeepseek30
deepseek-ai/deepseek-llm-67b-chatdeepseek30
deepseek-ai/DeepSeek-V2.5-1210deepseek30
deepseek-ai/DeepSeek-V3.1deepseek30
deepseek-ai/DeepSeek-V3deepseek33
deepseek-ai/DeepSeek-OCR-2deepseek20
deepseek-ai/DeepSeek-Prover-V2-7Bdeepseek20
deepseek-ai/deepseek-llm-67b-basedeepseek20
deepseek-ai/deepseek-coder-33b-basedeepseek20
deepseek-ai/DeepSeek-Prover-V1.5-Basedeepseek21
deepseek-ai/DeepSeek-Coder-V2-Instruct-0724deepseek20
deepseek-ai/deepseek-coder-5.7bmqa-basedeepseek20
deepseek-ai/DeepSeek-V3.2deepseek20
deepseek-ai/DeepSeek-V3.1-Terminusdeepseek27
deepseek-ai/DeepSeek-Prover-V2-671Bdeepseek20
deepseek-ai/DeepSeek-R1-Zerodeepseek25
deepseek-ai/DeepSeek-V2-Lite-Chatdeepseek18
deepseek-ai/deepseek-vl2-tinydeepseek10
deepseek-ai/DeepSeek-V2-Litedeepseek18
deepseek-ai/deepseek-moe-16b-basedeepseek10
deepseek-ai/deepseek-moe-16b-chatdeepseek10
deepseek-ai/deepseek-math-7b-instructdeepseek10
deepseek-ai/Janus-Pro-1Bdeepseek10
deepseek-ai/Janus-Pro-7Bdeepseek10
deepseek-ai/DeepSeek-V2-Chatdeepseek110
deepseek-ai/dspark_qwen3_4b_block7qwen10
deepseek-ai/dspark_qwen3_8b_block7qwen10
deepseek-ai/deepseek-vl2-smalldeepseek10
deepseek-ai/DeepSeek-Coder-V2-Instructdeepseek10
deepseek-ai/DeepSeek-V2-Chat-0628deepseek15
deepseek-ai/dspark_gemma4_12b_block7gemma10
deepseek-ai/deepseek-vl-1.3b-chatdeepseek10
deepseek-ai/dspark_qwen3_14b_block7qwen10
deepseek-ai/deepseek-vl-7b-chatdeepseek10
deepseek-ai/deepseek-math-7b-basedeepseek10
deepseek-ai/deepseek-vl2deepseek10
deepseek-ai/JanusFlow-1.3Bdeepseek10
deepseek-ai/dflash_qwen3_4b_block7qwen10
deepseek-ai/Janus-1.3Bdeepseek10
deepseek-ai/dflash_qwen3_8b_block7qwen10
deepseek-ai/dflash_gemma4_12b_block7gemma10
deepseek-ai/eagle3_qwen3_4b_ttt7qwen10
deepseek-ai/deepseek-math-7b-rldeepseek10
deepseek-ai/eagle3_qwen3_8b_ttt7qwen10
deepseek-ai/deepseek-coder-7b-base-v1.5deepseek10
deepseek-ai/eagle3_gemma4_12b_ttt7gemma10
deepseek-ai/dflash_qwen3_14b_block7qwen10
deepseek-ai/eagle3_qwen3_14b_ttt7qwen10
deepseek-ai/DeepSeek-Prover-V1.5-SFTdeepseek11
deepseek-ai/deepseek-vl-1.3b-basedeepseek10
deepseek-ai/deepseek-vl-7b-basedeepseek10
deepseek-ai/ESFT-vanilla-litedeepseek10
deepseek-ai/ESFT-token-intent-litedeepseek10
deepseek-ai/ESFT-token-law-litedeepseek10
deepseek-ai/ESFT-token-code-litedeepseek10
deepseek-ai/ESFT-token-math-litedeepseek10
deepseek-ai/ESFT-token-translation-litedeepseek10
deepseek-ai/ESFT-token-summary-litedeepseek10
deepseek-ai/DeepSeek-Prover-V1deepseek11
deepseek-ai/ESFT-gate-math-litedeepseek10
deepseek-ai/ESFT-gate-intent-litedeepseek10
deepseek-ai/ESFT-gate-code-litedeepseek10
deepseek-ai/ESFT-gate-translation-litedeepseek10
deepseek-ai/ESFT-gate-summary-litedeepseek10
deepseek-ai/ESFT-gate-law-litedeepseek10
deepseek-ai/DeepSeek-Math-V2deepseek10
deepseek-ai/DeepSeek-Prover-V1.5-RLdeepseek00

Top variants (101)

variantkindquantvRAMdlroot
DeepSeek-R1fp8fp88,860,456deepseek-ai/DeepSeek-R1
DeepSeek-V4-Flashfp8fp83,113,868deepseek-ai/DeepSeek-V4-Flash
DeepSeek-OCR-2originalbf162,607,706deepseek-ai/DeepSeek-OCR-2
DeepSeek-OCRoriginalbf162,569,577deepseek-ai/DeepSeek-OCR
DeepSeek-R1-0528-Qwen3-8Boriginalbf162,001,886deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
DeepSeek-V4-Profp8fp81,648,028deepseek-ai/DeepSeek-V4-Pro
DeepSeek-V3.2fp8fp81,268,192deepseek-ai/DeepSeek-V3.2-Exp-Base
DeepSeek-V2-Lite-Chatoriginalbf161,211,911deepseek-ai/DeepSeek-V2-Lite-Chat
DeepSeek-V3fp8fp81,159,954deepseek-ai/DeepSeek-V2
deepseek-vl2-tinyoriginalbf161,121,071deepseek-ai/deepseek-vl2-tiny
DeepSeek-V3-0324fp8fp8835,522deepseek-ai/DeepSeek-V3-0324
DeepSeek-R1-Distill-Qwen-32Boriginalbf16832,252deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
DeepSeek-R1-Distill-Llama-70Boriginalbf16778,263deepseek-ai/DeepSeek-R1-Distill-Llama-70B
DeepSeek-R1-Distill-Qwen-1.5Boriginalbf16607,200deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
DeepSeek-Coder-V2-Lite-Instructoriginalbf16534,008deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
DeepSeek-R1-Distill-Qwen-14Boriginalbf16504,193deepseek-ai/DeepSeek-R1-Distill-Qwen-14B
DeepSeek-V2-Liteoriginalbf16499,750deepseek-ai/DeepSeek-V2-Lite
deepseek-coder-6.7b-instructoriginalbf16472,346deepseek-ai/deepseek-coder-6.7b-instruct
deepseek-coder-7b-instruct-v1.5originalbf16469,383deepseek-ai/deepseek-coder-7b-instruct-v1.5
DeepSeek-R1-Distill-Llama-8Boriginalbf16408,725deepseek-ai/DeepSeek-R1-Distill-Llama-8B
DeepSeek-V4-Flash-DSparkfp8fp8394,098—
DeepSeek-V3.2-Expfp8fp8291,271deepseek-ai/DeepSeek-V3.2-Exp-Base
DeepSeek-V3.1fp8fp8287,286deepseek-ai/DeepSeek-V3.1-Base
DeepSeek-R1-Distill-Qwen-7Boriginalbf16280,986deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
DeepSeek-R1-0528fp8fp8247,374deepseek-ai/DeepSeek-R1-0528
DeepSeek-V4-Flash-Basefp8fp8205,182deepseek-ai/DeepSeek-V4-Flash
deepseek-coder-1.3b-instructoriginalbf1672,165deepseek-ai/deepseek-coder-1.3b-instruct
deepseek-coder-6.7b-baseoriginalbf1658,464deepseek-ai/deepseek-coder-6.7b-base
DeepSeek-V4-Pro-Basefp8fp846,096deepseek-ai/DeepSeek-V4-Pro
DeepSeek-V4-Pro-DSparkfp8fp842,123—
deepseek-llm-7b-chatoriginalbf1641,614deepseek-ai/deepseek-llm-7b-chat
deepseek-moe-16b-baseoriginalbf1635,504deepseek-ai/deepseek-moe-16b-base
deepseek-llm-7b-baseoriginalbf1635,501deepseek-ai/deepseek-llm-7b-base
DeepSeek-V3-Basefp8fp834,085deepseek-ai/DeepSeek-V2
DeepSeek-Prover-V2-7Boriginalbf1628,335deepseek-ai/DeepSeek-Prover-V2-7B
DeepSeek-V3.1-Basefp8fp824,203deepseek-ai/DeepSeek-V3.1-Base
deepseek-moe-16b-chatoriginalbf1618,044deepseek-ai/deepseek-moe-16b-chat
DeepSeek-V3.1-Terminusfp8fp814,951deepseek-ai/DeepSeek-V3.1-Base
deepseek-math-7b-instructoriginalbf1612,977deepseek-ai/deepseek-math-7b-instruct
Janus-Pro-1Boriginalbf1612,147deepseek-ai/Janus-Pro-1B
Janus-Pro-7Boriginalbf1611,650deepseek-ai/Janus-Pro-7B
deepseek-coder-1.3b-baseoriginalbf1610,987deepseek-ai/deepseek-coder-1.3b-base
DeepSeek-V2-Chatoriginalbf1610,482deepseek-ai/DeepSeek-V2-Chat
dspark_qwen3_4b_block7originalbf1610,060deepseek-ai/dspark_qwen3_4b_block7
DeepSeek-V2originalbf169,375deepseek-ai/DeepSeek-V2
deepseek-llm-67b-baseoriginalbf168,938deepseek-ai/deepseek-llm-67b-base
dspark_qwen3_8b_block7originalbf166,462deepseek-ai/dspark_qwen3_8b_block7
deepseek-vl2-smalloriginalbf165,661deepseek-ai/deepseek-vl2-small
DeepSeek-Coder-V2-Instructoriginalbf165,369deepseek-ai/DeepSeek-Coder-V2-Base
DeepSeek-R1-Zerofp8fp85,179deepseek-ai/DeepSeek-R1-Zero
DeepSeek-Coder-V2-Lite-Baseoriginalbf165,001deepseek-ai/DeepSeek-Coder-V2-Lite-Base
DeepSeek-V2.5originalbf164,430deepseek-ai/DeepSeek-V2.5
deepseek-coder-33b-instructoriginalbf163,772deepseek-ai/deepseek-coder-33b-instruct
DeepSeek-V2-Chat-0628originalbf163,439deepseek-ai/DeepSeek-V2-Chat-0628
dspark_gemma4_12b_block7originalbf163,271deepseek-ai/dspark_gemma4_12b_block7
deepseek-vl-1.3b-chatoriginalbf163,269deepseek-ai/deepseek-vl-1.3b-chat
dspark_qwen3_14b_block7originalbf162,998deepseek-ai/dspark_qwen3_14b_block7
deepseek-coder-33b-baseoriginalbf162,720deepseek-ai/deepseek-coder-33b-base
deepseek-vl-7b-chatoriginalbf162,703deepseek-ai/deepseek-vl-7b-chat
deepseek-math-7b-baseoriginalbf162,573deepseek-ai/deepseek-math-7b-base
deepseek-vl2originalbf162,399deepseek-ai/deepseek-vl2
JanusFlow-1.3Boriginalbf162,317deepseek-ai/JanusFlow-1.3B
DeepSeek-V3.2-Specialefp8fp82,254deepseek-ai/DeepSeek-V3.2-Exp-Base
dflash_qwen3_4b_block7originalbf162,212deepseek-ai/dflash_qwen3_4b_block7
DeepSeek-Coder-V2-Baseoriginalbf161,809deepseek-ai/DeepSeek-Coder-V2-Base
DeepSeek-Prover-V1.5-RLoriginalbf161,782deepseek-ai/DeepSeek-Prover-V1.5-SFT
Janus-1.3Boriginalbf161,636deepseek-ai/Janus-1.3B
dflash_qwen3_8b_block7originalbf161,562deepseek-ai/dflash_qwen3_8b_block7
dflash_gemma4_12b_block7originalbf161,103deepseek-ai/dflash_gemma4_12b_block7
deepseek-llm-67b-chatoriginalbf161,034deepseek-ai/deepseek-llm-67b-chat
eagle3_qwen3_4b_ttt7originalbf161,029deepseek-ai/eagle3_qwen3_4b_ttt7
deepseek-math-7b-rloriginalbf161,000deepseek-ai/deepseek-math-7b-rl
eagle3_qwen3_8b_ttt7originalbf16978deepseek-ai/eagle3_qwen3_8b_ttt7
deepseek-coder-7b-base-v1.5originalbf16852deepseek-ai/deepseek-coder-7b-base-v1.5
eagle3_gemma4_12b_ttt7originalbf16803deepseek-ai/eagle3_gemma4_12b_ttt7
dflash_qwen3_14b_block7originalbf16735deepseek-ai/dflash_qwen3_14b_block7
DeepSeek-V2.5-1210originalbf16725deepseek-ai/DeepSeek-V2.5-1210
eagle3_qwen3_14b_ttt7originalbf16689deepseek-ai/eagle3_qwen3_14b_ttt7
DeepSeek-Prover-V2-671Bfp8fp8641deepseek-ai/DeepSeek-Prover-V2-671B
DeepSeek-Math-V2fp8fp8427deepseek-ai/DeepSeek-Math-V2
DeepSeek-Prover-V1.5-Baseoriginalbf16424deepseek-ai/DeepSeek-Prover-V1.5-Base
DeepSeek-Coder-V2-Instruct-0724originalbf16407deepseek-ai/DeepSeek-Coder-V2-Base
DeepSeek-Prover-V1.5-SFToriginalbf16392deepseek-ai/DeepSeek-Prover-V1.5-Base
DeepSeek-V3.2-Exp-Basefp8fp8307deepseek-ai/DeepSeek-V3.2-Exp-Base
deepseek-coder-5.7bmqa-baseoriginalbf16294deepseek-ai/deepseek-coder-5.7bmqa-base
deepseek-vl-1.3b-baseoriginalbf16147deepseek-ai/deepseek-vl-1.3b-base
deepseek-vl-7b-baseoriginalbf16122deepseek-ai/deepseek-vl-7b-base
ESFT-vanilla-liteoriginalbf16117deepseek-ai/ESFT-vanilla-lite
ESFT-token-intent-liteoriginalbf16102deepseek-ai/ESFT-token-intent-lite
ESFT-token-law-liteoriginalbf1696deepseek-ai/ESFT-token-law-lite
ESFT-token-code-liteoriginalbf1695deepseek-ai/ESFT-token-code-lite
ESFT-token-math-liteoriginalbf1694deepseek-ai/ESFT-token-math-lite
ESFT-token-translation-liteoriginalbf1692deepseek-ai/ESFT-token-translation-lite
ESFT-token-summary-liteoriginalbf1691deepseek-ai/ESFT-token-summary-lite
DeepSeek-Prover-V1originalbf1682deepseek-ai/DeepSeek-Prover-V1
ESFT-gate-math-liteoriginalbf1679deepseek-ai/ESFT-gate-math-lite
ESFT-gate-intent-liteoriginalbf1677deepseek-ai/ESFT-gate-intent-lite
ESFT-gate-code-liteoriginalbf1675deepseek-ai/ESFT-gate-code-lite
ESFT-gate-translation-liteoriginalbf1672deepseek-ai/ESFT-gate-translation-lite
ESFT-gate-summary-liteoriginalbf1671deepseek-ai/ESFT-gate-summary-lite
ESFT-gate-law-liteoriginalbf1668deepseek-ai/ESFT-gate-law-lite
Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.