Skip to content

What can the 2× RTX 4090 run?

What fits on the 2× RTX 4090
Text models: 188Visual models: 0Strongest first
Memory
48 GB
Bandwidth
1,008 GB/s each
Maker
NVIDIA

Open the calculator on this machine

Text models that run on it

Text models that run on the 2× RTX 4090, best first
ModelBest fitContextSpeedStrongest at
Qwen3.8 27B8-bit256K24 tok/sChat, Coding, Agents, Docs
Muse Glimmer 30B8-bit128K23 tok/sDocs
Nemotron 3 Super 120B A12B2-bit96K84 tok/sChat, Coding, Agents, Docs
Nemotron 3.5 Bolt 74B A9B4-bit216K75 tok/sChat, Coding, Agents, Docs
Qwen3.6 35B A3B8-bit256K143 tok/sChat, Coding, Agents, Docs
Gemma 4 31B8-bit256K20 tok/sChat, Coding, Agents, Docs
Gemma 4 26B A4B8-bit256K133 tok/sChat, Coding, Agents, Docs
Qwen3.5 9B8-bit256K71 tok/sChat, Coding, Agents, Docs
GLM-4.7 Flash8-bit198K136 tok/sChat, Coding, Agents
Gemma 4 12B8-bit256K51 tok/sChat, Coding, Docs
Mistral Small 4 119B2-bit104K178 tok/sChat, Docs
gpt-oss 120B2-bit88K205 tok/sChat, Coding
Qwen3 Coder Next 80B3-bit256K197 tok/sCoding, Agents
Nemotron 3.5 Lightning 30B A3B8-bit256K130 tok/sAgents, Docs
Devstral Small 2 24BFP8128K27 tok/sCoding
LFM2.5 24B A2B8-bit128K185 tok/sAgents
Mistral Small 3.2 24B8-bit128K28 tok/sChat, Coding, Agents
Ministral 3 14B8-bit192K47 tok/sCoding, Agents, Docs
gpt-oss 20BMXFP4128K197 tok/sAgents
Ministral 3 8B8-bit256K73 tok/sDocs
Llama 3.3 70B3-bit36K20 tok/sChat, Coding, Agents, Docs
Olmo 3.1 32B8-bit64K20 tok/sChat, Coding, Agents, Docs
Llama 3.1 8B8-bit128K82 tok/sChat, Coding, Agents, Docs
Qwen3 14B8-bit40K45 tok/sNot rated
Qwen3 VL 8B Instruct8-bit248K75 tok/sText and images
Qwen3 8B8-bit40K80 tok/sNot rated
Qwen3 32B8-bit40K20 tok/sNot rated
Qwen3.6 27B8-bit256K24 tok/sText and images
Qwen2.5 Coder 14B Instruct8-bit32K44 tok/sCoding
Qwen3 Coder 30B A3B Instruct8-bit144K135 tok/sCoding
Qwen3.5 27B8-bit256K24 tok/sText and images
Qwen3.5 35B A3B8-bit256K143 tok/sText and images
NVIDIA Nemotron 3 Nano 30B A3B BF168-bit256K130 tok/sNot rated
Mistral 7B Instruct V0.28-bit32K86 tok/sNot rated
GLM 4.5 Air2-bit28K93 tok/sNot rated
Qwen3 30B A3B8-bit40K135 tok/sNot rated
Qwen3 30B A3B Instruct 25078-bit144K135 tok/sNot rated
Laguna XS.28-bit256K144 tok/sNot rated
OTel 2.0 LLM 31B IT8-bit256K20 tok/sNot rated
Qwen2.5 32B Instruct8-bit32K20 tok/sNot rated
MiniCPM V 4.7 35B A3B8-bit256K143 tok/sText and images
Qwen2.5 14B Instruct8-bit32K44 tok/sNot rated
Laguna XS 2.18-bit256K144 tok/sNot rated
Diffusiongemma 26B A4B IT8-bit256K128 tok/sText and images
Qwen2.5 VL 32B Instruct8-bit40K20 tok/sText and images
Qwen2.5 Coder 32B Instruct8-bit32K20 tok/sCoding
Laguna S 2.12-bit64K136 tok/sNot rated
Ornith 1.5 9B8-bit256K71 tok/sNot rated
Qwen3 VL 30B A3B Instruct8-bit136K135 tok/sText and images
EXAONE 3.5 7.8B Instruct8-bit32K83 tok/sNot rated
DeepSeek R1 0528 Qwen3 8B8-bit128K80 tok/sNot rated
Gemma 3 12B IT8-bit128K55 tok/sText and images
Phi 3.5 Mini Instruct8-bit104K152 tok/sNot rated
DeepSeek R1 Distill Qwen 32B8-bit44K20 tok/sNot rated
Gemma 3 27B IT8-bit128K24 tok/sText and images
Ornith 1.5 35B A3B8-bit256K143 tok/sNot rated
GPT Neox 20B8-bit2K31 tok/sNot rated
Apertus 8B Instruct 25098-bit64K81 tok/sNot rated
Phi 48-bit16K44 tok/sNot rated
Phi 3 Mini 128K Instruct8-bit104K152 tok/sNot rated

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 188 are shown.

Only just too big

These need up to a quarter more memory than the 2× RTX 4090 has. Spilling into system RAM or a lower precision can get them running, slowly.

Questions

What is the best LLM for the 2× RTX 4090?

Of the open models we track, the strongest that run well on the 2× RTX 4090 are Qwen3.8 27B, Muse Glimmer 30B, Nemotron 3 Super 120B A12B. 188 text models fit it in all.

Calculated from published specs, not benchmarked. All machines