Graphics cards
Which AI models run on the RX 7900 XT (20 GB)?
The RX 7900 XT (20 GB) has 20 GB of memory, about 19.0 GB of it usable for a model, and 800 GB/s of memory bandwidth. 63 of the 110 models in our catalog run on it at 4-bit or better.
Data as of 2 November 2026
20 GBmemory
~19.0 GB usable for a model
800 GB/sbandwidth, sets the speed
63models run at 4-bit or better
Models that run on it
| Model | Best build | Memory | Speed |
|---|---|---|---|
| Qwen3.8-27B27.3B | Q4_K_M | 17.9 GB | ~37 tok/s ModelFit est. |
| Gemma 4 26B-A4B25.2B MoE | Q4_K_M | 17.4 GB | ~111 tok/s ModelFit est. |
| gpt-oss-20b21B MoE | MXFP4 | 15.1 GB | ~112 tok/s ModelFit est. |
| Nemotron Nano 9B v29B | Q8_0 | 10.8 GB | ~46 tok/s our est. |
| Qwen3.5 9B Instruct9B | Q8_0 | 12.0 GB | ~65 tok/s ModelFit est. |
| Qwen2.5 7B Instruct7B | Q4_K_M | 6.6 GB | ~103 tok/s ModelFit est. |
| Gemma 3 4B Instruct4B | Q8_0 | 5.0 GB | ~107 tok/s ModelFit est. |
| Muse Glimmer 30B29.8B | Q4_K_M | 18.3 GB | ~37 tok/s ModelFit est. |
| Granite 4.2 30B29.3B | Q4_K_M | 17.9 GB | no estimate |
| AREX-2 27B27.4B | Q4_K_M | 18.6 GB | ~36 tok/s ModelFit est. |
| Qwen3.5 27B Instruct27B | Q4_K_M | 17.4 GB | ~38 tok/s ModelFit est. |
| Jebadiah 27B27B | Q4_K_M | 18.2 GB | ~37 tok/s ModelFit est. |
| Mistral Small 3.124B | Q4_K_M | 16.4 GB | ~29 tok/s our est. |
| LFM2 24B-A2B Instruct24B MoE | Q4_K_M | 15.3 GB | ~160 tok/s ModelFit est. |
| Devstral Small 2 24B24B | Q4_K_M | 16.6 GB | ~39 tok/s ModelFit est. |
| Mistral Small 22B22B | Q4_K_M | 18.4 GB | ~44 tok/s ModelFit est. |
| Qwen2.5 14B Instruct14B | Q4_K_M | 12.3 GB | ~60 tok/s ModelFit est. |
| Phi-3 Medium 14B14B | Q4_K_M | 12.3 GB | ~40 tok/s our est. |
| Phi-4 14B14B | Q8_0 | 15.9 GB | ~38 tok/s ModelFit est. |
| Qwen2.5 Coder 14B14B | Q4_K_M | 12.3 GB | ~60 tok/s ModelFit est. |
| Qwen3 14B14B | Q8_0 | 17.3 GB | ~38 tok/s ModelFit est. |
| DeepSeek-R1 Distill Qwen 14B14B | Q8_0 | 16.0 GB | ~38 tok/s ModelFit est. |
| Ministral 3 14B13.5B | Q4_K_M | 10.3 GB | ~60 tok/s ModelFit est. |
| Mellum2.1 12B-A2.5B Thinking12.1B MoE | Q4_K_M | 9.3 GB | ~263 tok/s our est. |
| Mistral Nemo 12B12B | Q8_0 | 13.4 GB | ~45 tok/s ModelFit est. |
| Gemma 3 12B Instruct12B | Q4_K_M | 10.7 GB | ~68 tok/s ModelFit est. |
| Gemma 4 12B12B | Q8_0 | 14.1 GB | ~44 tok/s ModelFit est. |
| GLM-4-9B-Chat9.4B | Q8_0 | 11.2 GB | ~59 tok/s ModelFit est. |
| Gemma 2 9B Instruct9B | Q4_K_M | 8.2 GB | ~73 tok/s ModelFit est. |
| Ornith 1.0 9B9B | Q4_K_M | 6.7 GB | ~96 tok/s ModelFit est. |
| MiMo-V2.6-Distill-Qwen-9B9B | Q4_K_M | 6.9 GB | ~93 tok/s ModelFit est. |
| Jebadiah 9B9B | Q8_0 | 11.0 GB | ~63 tok/s ModelFit est. |
| Granite 4.2 8B8.8B | Q4_K_M | 6.1 GB | ~87 tok/s ModelFit est. |
| Granite Switch 4.1 8B8.8B | Q4_K_M | 7.6 GB | ~74 tok/s ModelFit est. |
| MiniCPM-V 4.5 8B8.7B | Q4_K_M | 6.8 GB | ~93 tok/s ModelFit est. |
| Ministral 3 8B8.4B | Q4_K_M | 7.1 GB | ~82 tok/s ModelFit est. |
| LFM2.5 8B-A1B8.3B MoE | Q4_K_M | 6.6 GB | ~186 tok/s ModelFit est. |
| Llama 3.1 8B Instruct8B | Q8_0 | 9.2 GB | ~64 tok/s ModelFit est. |
| Qwen3 8B8B | Q8_0 | 9.3 GB | ~63 tok/s ModelFit est. |
| Granite 4.1 8B Instruct8B | Q4_K_M | 6.6 GB | ~84 tok/s ModelFit est. |
| Mistral 7B Instruct7B | Q8_0 | 8.4 GB | ~67 tok/s ModelFit est. |
| Qwen2.5 Coder 7B7B | Q4_K_M | 6.6 GB | ~103 tok/s ModelFit est. |
| DeepSeek-R1 Distill Qwen 7B7B | Q8_0 | 8.7 GB | ~69 tok/s ModelFit est. |
| Gemma 4 E4B4.5B | Q8_0 | 8.7 GB | ~96 tok/s ModelFit est. |
| Qwen3.5 4B Instruct4B | Q8_0 | 5.4 GB | ~102 tok/s ModelFit est. |
| NVIDIA Nemotron 3.5 Bolt E4B4B | Q4_K_M | 9.1 GB | ~115 tok/s ModelFit est. |
| Phi-3 Mini 3.8B3.8B | Q4_K_M | 4.3 GB | ~138 tok/s our est. |
| Phi-4 Mini 3.8B3.8B | Q8_0 | 4.9 GB | ~102 tok/s ModelFit est. |
| Granite 4.2 3B3.7B | Q4_K_M | 3.2 GB | ~153 tok/s ModelFit est. |
| Ministral 3 3B3.4B | Q4_K_M | 4.1 GB | ~126 tok/s ModelFit est. |
| Qwen2.5 3B Instruct3B | Q4_K_M | 3.6 GB | ~167 tok/s ModelFit est. |
| Llama 3.2 3B Instruct3B | Q4_K_M | 3.6 GB | ~151 tok/s ModelFit est. |
| Granite 4.1 3B Instruct3B | Q4_K_M | 3.0 GB | ~153 tok/s ModelFit est. |
| Gemma 4 E2B2.3B | Q8_0 | 5.7 GB | ~166 tok/s ModelFit est. |
| Gemma 2 2B Instruct2B | Q4_K_M | 2.8 GB | ~163 tok/s ModelFit est. |
| Qwen3.5 2B Instruct2B | Q4_K_M | 2.8 GB | ~167 tok/s ModelFit est. |
| LFM3 1.7B1.73B | Q4_K_M | 2.1 GB | ~216 tok/s ModelFit est. |
| Qwen2.5 1.5B Instruct1.5B | Q4_K_M | 2.5 GB | ~220 tok/s ModelFit est. |
| Llama 3.2 1B Instruct1B | Q4_K_M | 2.0 GB | ~232 tok/s ModelFit est. |
| Gemma 3 1B Instruct1B | Q8_0 | 2.0 GB | ~220 tok/s ModelFit est. |
| Qwen3.5 0.8B Instruct0.8B | Q4_K_M | 1.8 GB | ~239 tok/s ModelFit est. |
| Qwen2.5 0.5B Instruct0.5B | Q4_K_M | 1.8 GB | ~280 tok/s ModelFit est. |
| SmolLM2 360M0.36B | Q4_K_M | 1.5 GB | ~236 tok/s ModelFit est. |
Speeds come from ModelFit's estimates (CC BY 4.0) where it covers the model, otherwise from our bandwidth formula. Method
Fits only at 3-bit or lower
Too big for this machine
Qwen3.6 27B (19.4 GB), Qwen3-Coder-30B-A3B (20 GB), Nemotron 3.5 Lightning 30B-A3B (20 GB), Cohere North Mini Code (20 GB), GLM-4.7-Flash 30B-A3B (20 GB), Olmo 3.1 32B Instruct (21 GB), DeepSeek-R1 Distill Qwen 32B (21 GB), Gemma 4 31B (21 GB), Qwen3.5 35B-A3B Instruct (21 GB), Qwen3 32B (22 GB), Laguna XS 2.1 (22 GB), Nanbeige4.3 33B (22 GB) and more.
Questions
What is the best AI model for the RX 7900 XT (20 GB)?
Qwen3.8-27B at Q4_K_M: the largest model on our shortlist of current models that runs at 4-bit or better at a usable speed (about 37 tok/s, ModelFit est.).
How much memory can a model use on the RX 7900 XT (20 GB)?
About 19.0 GB of the 20 GB.
How fast is the RX 7900 XT (20 GB) for local LLMs?
Decode speed is limited by memory bandwidth: 800 GB/s. A model that reads 10 GB per token runs at under 80 tokens per second.