Skip to content

What can the RTX 4080 Super run?

What fits on the RTX 4080 Super
Text models: 247Visual models: 18Strongest first
Memory
16 GB
Bandwidth
736 GB/s
Maker
NVIDIA

Open the calculator on this machineEvery model for 16 GB of VRAM

Text models that run on it

Text models that run on the RTX 4080 Super, best first
ModelBest fitContextSpeedStrongest at
Qwen3.8 27B3-bit20K37 tok/sChat, Coding, Agents, Docs
Muse Glimmer 30B3-bit40K35 tok/sDocs
Gemma 4 26B A4B3-bit224K185 tok/sChat, Coding, Agents, Docs
Qwen3.5 9B8-bit144K53 tok/sChat, Coding, Agents, Docs
Gemma 4 12B8-bit248K38 tok/sChat, Coding, Docs
Qwen3.5 4B8-bit256K94 tok/sAgents, Docs
Devstral Small 2 24B3-bit20K43 tok/sCoding
LFM2.5 24B A2B4-bit44K210 tok/sAgents
Mistral Small 3.2 24B3-bit20K43 tok/sChat, Coding, Agents
Ministral 3 14B6-bit20K45 tok/sCoding, Agents, Docs
Gemma 4 E4B8-bit128K84 tok/sChat, Coding, Docs
gpt-oss 20BMXFP4120K156 tok/sAgents
Ministral 3 8B8-bit40K55 tok/sDocs
Gemma 4 E2B8-bit128K148 tok/sChat
Nemotron 3 Nano 4B8-bit256K114 tok/sDocs
Llama 3.1 8B8-bit48K61 tok/sChat, Coding, Agents, Docs
Qwen3 0.6B8-bit40K418 tok/sNot rated
Qwen3 14B6-bit16K43 tok/sNot rated
Qwen3 VL 8B Instruct8-bit40K56 tok/sText and images
Qwen3 8B8-bit40K60 tok/sNot rated
Qwen2.5 7B Instruct8-bit32K65 tok/sNot rated
Qwen2.5 VL 7B Instruct8-bit112K59 tok/sText and images
Unlimited OCR8-bit32K325 tok/sReading documents
Qwen3 VL 4B Instruct8-bit64K98 tok/sText and images
Qwen3 1.7B8-bit40K199 tok/sNot rated
Qwen3 4B8-bit40K107 tok/sNot rated
Qwen3 4B Instruct 25078-bit72K107 tok/sNot rated
Qwen2.5 3B Instruct8-bit32K134 tok/sNot rated
Qwen3.6 27B3-bit20K37 tok/sText and images
DeepSeek OCR8-bit8K325 tok/sReading documents
Qwen2.5 Coder 14B Instruct6-bit16K42 tok/sCoding
Qwen2.5 VL 3B Instruct8-bit125K113 tok/sText and images
Qwen3.5 27B3-bit20K37 tok/sText and images
Qwen2.5 Coder 7B Instruct8-bit32K65 tok/sCoding
Mistral 7B Instruct V0.28-bit32K64 tok/sNot rated
Chandra OCR 28-bit256K83 tok/sReading documents
Qwen3 VL 2B Instruct8-bit112K191 tok/sText and images
MiniCPM5 2B8-bit128K170 tok/sNot rated
Gemma 3 4B IT8-bit128K118 tok/sText and images
Llama 3.2 3B Instruct8-bit96K134 tok/sNot rated
Qwen2.5 14B Instruct6-bit16K42 tok/sNot rated
Qwen3 1.7B Base8-bit32K228 tok/sNot rated
Diffusiongemma 26B A4B IT3-bit112K179 tok/sText and images
HunyuanOCR8-bit128K330 tok/sText and images
Qwen3 0.6B Base8-bit32K483 tok/sNot rated
Qwen2 VL 7B Instruct8-bit32K59 tok/sText and images
Ornith 1.5 9B8-bit144K53 tok/sNot rated
Unlimited OCR 28-bit128K324 tok/sReading documents
Granite 4.1 3B8-bit128K122 tok/sNot rated
Mistral 7B V0.18-bit32K64 tok/sNot rated
Qwen2.5 Coder 3B Instruct8-bit32K134 tok/sCoding
Meta Llama 3 8B Instruct Abliterated V38-bit8K61 tok/sNot rated
EXAONE 3.5 7.8B Instruct8-bit32K62 tok/sNot rated
Saiga Llama3 8B8-bit8K61 tok/sNot rated
DeepSeek R1 0528 Qwen3 8B8-bit44K60 tok/sNot rated
Gemma 3 12B IT8-bit28K41 tok/sText and images
Phi 3.5 Mini Instruct8-bit28K116 tok/sNot rated
Qwen3 4B Thinking 25078-bit72K107 tok/sNot rated
Phi 4 Mini Instruct8-bit80K113 tok/sNot rated
Dots.ocr8-bit128K151 tok/sReading documents

Best fit is the highest quality, up to 8-bit, that leaves the use case’s context. Ranked by our use-case scores, then by Hugging Face downloads. Speeds are rough, plus or minus 25%. The top 60 of 247 are shown.

Image and video models

Image and video models that run on the RTX 4080 Super
ModelBest fitPer outputStrongest at
LTX-2.55-bit82 sVideo
Z-Image Turbo16-bit9.5 sImages
FLUX.1 [dev]8-bit18 sImages
Qwen-Image5-bit6.3 minImages
Qwen-Image 2.116-bit2.4 minImages
FLUX.3 Klein16-bit14 sImages
Wan2.2 I2V A14B4-bit56 minVideo
FLUX.2 [klein] 9B8-bit4.9 sImages
Wan2.1 I2V 14B4-bit56 minVideo
Wan 2.1 T2V-14B4-bit70 minVideo
Krea 2 Turbo8-bit18 sImages
Stable Diffusion 3.5 Large8-bit75 sImages
Qwen Image 25125-bit6.3 minImages
MOSS-Image16-bit1.6 minImages
Wan 2.2 TI2V-5B16-bit13 minVideo
MOSS Image Base16-bit1.6 minImages
Wan 2.2 A14B4-bit56 minVideo
HunyuanVideo 1.58-bit60 minVideo

Only just too big

Questions

What is the best LLM for the RTX 4080 Super?

Of the open models we track, the strongest that run well on the RTX 4080 Super are Qwen3.8 27B, Muse Glimmer 30B, Gemma 4 26B A4B. 247 text models fit it in all.

Calculated from published specs, not benchmarked. All machines