Best Local AI Models for AMD RX 7900 XT (20GB)

The RX 7900 XT sits in a capacity gap NVIDIA does not serve: 20GB of VRAM, more than any 16GB card and cheaper than any 24GB one. That extra 4GB over a 5070 Ti lets 27B-class models load fully in VRAM at Q4, and 800 GB/s of bandwidth keeps generation quick. Like the XTX, it runs Ollama and llama.cpp through ROCm with a Vulkan fallback.

Compare with another device
Quick answer

The best local LLM for the RX 7900 XT is Qwen3.8 27B (est. 37 tok/s, typical error 50%) on its 20GB VRAM. It uses ~16.5GB of VRAM; the RX 7900 XT handles dense models up to ~30B at Q4. A 14B model at Q4 runs at est. 60 tok/s, typical error 10%. (ModelFit, 2026)

Sizing rule: a Q4 model needs about 0.6 GB of VRAM per billion parameters for its weights, and ModelFit budgets 90% of this card's 20GB (18GB) for weights, context, and KV-cache. Real files often load more than that rule, so the per-size table below checks actual builds: a size fits only when the lightest Q4 build of that size in our catalog loads within 18GB. MoE rows, and sizes with no such build, use a typical Q4 file size. Other strong fits: Qwen3.5 9B Instruct (Q8) (9B, ~10.7GB) and Gemma 4 12B (Q8) (12B, ~12.8GB). On this card Qwen3.8 27B runs at est. 37 tok/s, typical error 50%. What it will not run: a 32B model does not fit in VRAM alone, so it spills to system RAM over PCIe and slows sharply.

$ollama run qwen3.8:27b
Top pick
Qwen3.8 27B
Decode speed
est. 37 tok/s, typical error 50%
VRAM needed
~16.5 GB

Speeds are ModelFit estimates from memory bandwidth and model size, not measured benchmarks. Typical error: the median gap between the ModelFit speed model and measured runs on hardware the model was not fitted on, for this backend and model size.

Cite this page: ModelFit, RX 7900 XT 20GB Local LLM: Runs 27B at est. 37 tok/s, https://modelfit.io/gpu/rx-7900-xt/, updated November 2026, CC BY 4.0.

Specs
VRAM
20 GB GDDR6
Speed (8B Q4)
est. 91 tok/stypical error 10%
Bandwidth
800 GB/s
Architecture
RDNA 3
Max model size
Up to ~30B dense
Compatibility
10 excellent, 0 workable
Table 1

RX 7900 XT estimated tokens/sec by model size

Q4_K_M · ModelFit estimate
Model sizeEst. speedFit on 20GB
7Best. 102 tok/stypical error 10%Fits in VRAM
14Best. 60 tok/stypical error 10%Fits in VRAM
20B MoE (3.6B active)est. 114 tok/stypical error 25%Fits in VRAM
32Best. 6 tok/srough estimate, offload not measuredCPU offload (slow)
35B MoE (3B active)est. 96 tok/srough estimate, offload not measuredCPU offload (slow)
70Best. 1 tok/srough estimate, offload not measuredCPU offload (slow)
120B MoE (5.1B active)est. 31 tok/srough estimate, offload not measuredCPU offload (slow)

ModelFit estimates, not measured benchmarks: decode speed of a generic Q4_K_M model of each size at 4K context on the RX 7900 XT's 800 GB/s bandwidth. MoE rows scale by active parameters (decode reads only the active experts), so a 35B MoE runs far faster than a dense 32B. "CPU offload" sizes exceed the 20GB VRAM; dense models slow to a crawl there, MoE models degrade less because hot experts stay GPU-resident. Offload rows are rough estimates: no measured run covers offload, so they carry no typical error. What typical error means.

Fig. 2

Context costs VRAM too.

Qwen3.8 27B loads ~16.5 GB of weights; at 16k context the KV cache adds ~1.0 GB (still fits the ~18 GB usable VRAM), and at 64k it adds ~4.0 GB (exceeds the budget, use a smaller quant or a q8_0 KV cache).

Weights 17 GBKV cache · 16k context 1.0 GBHeadroom 0.5 GBRuntime reserve 2.0 GB

The model leaves about 0.5 GB of the usable VRAM budget free after weights and 16k context.

KV-cache figures assume an fp16 cache, the llama.cpp/Ollama default. Standard GQA models use a size-class estimate (8 KV heads x 128 head dim class); hybrid linear-attention models (Qwen3.5/3.6, Qwen3-Next) use the exact per-token cost from their published config, since only their sparse full-attention layers cache KV. A q8_0 KV cache roughly halves either figure. Estimates, not measurements.

RX 7900 XT VRAM for AI: what actually fits?

20GB GDDR6 at 800 GB/s is a unique spot in the lineup: 16GB cards top out at 14B-class dense models, while the 7900 XT fits a 27B at Q4 (~18GB) right at its usable budget, with the KV cache limiting context length. It cannot hold 32B Q4 (~20GB), so treat 27B as the practical ceiling. RDNA 3 is on the official ROCm support list, so Ollama, llama.cpp, and LM Studio all run it without CUDA. On the used market it usually undercuts every 24GB card, making it a strong value pick for the 14B-27B class.

Table 2

RX 7900 XT vs top GPUs

HardwareMemorySpeedBandwidth
RTX 5070 Ti16 GBest. 125 tok/stypical error 5%896 GB/s
RX 7900 XT20 GBest. 91 tok/stypical error 10%800 GB/s
RX 7900 XTX24 GBest. 104 tok/stypical error 10%960 GB/s
RTX 309024 GBest. 117 tok/stypical error 5%936 GB/s

Recommended models

registry-verified10 models
01

Qwen3.8 27B

Qwen / 27B / Q4_K_M / ~16.5 GB

Local OKOKB · 8

Best for: Coding, Agent, Vision, Long context·Pop: 95/100

Perf: est. 37 tok/s, typical error 50% · first token ~1.0s

Fits in 20 GB VRAM with room to spare. Best for coding, agent, vision, long context on RX 7900 XT.

ollamaregistry-verified
02

Qwen3.5 9B Instruct (Q8)

Qwen / 9B / Q8_0 / ~10.7 GB

Local OKExcellentA · 41

Best for: Quality, Coding, Reasoning·Pop: 86/100

Perf: est. 65 tok/s, typical error 10% · first token ~1.7s

Fits in 20 GB VRAM with room to spare. Best for quality, coding, reasoning on RX 7900 XT.

ollamaregistry-verified
03

Gemma 4 12B (Q8)

Gemma / 12B / Q8_0 / ~12.8 GB

Local OKOKB · 29

Best for: Chat, Coding, Multimodal·Pop: 80/100

Perf: est. 44 tok/s, typical error 10% · first token ~2.4s

Fits in 20 GB VRAM with room to spare. Best for chat, coding, multimodal on RX 7900 XT.

ollamaregistry-verified
04

Gemma 4 26B-A4B

Gemma / 26B / Q4_K_M / ~16 GB

Local OKOKB · 11

Best for: Chat, Coding, Multimodal·Pop: 86/100

Perf: est. 111 tok/s, typical error 25% · first token ~0.3s

Fits in 20 GB VRAM with room to spare. Best for chat, coding, multimodal on RX 7900 XT.

ollamaregistry-verified
05

LFM2 24B-A2B Instruct

LFM2 / 24B / Q4_K_M / ~14 GB

Local OKOKB · 22

Best for: Local AI agents, privacy-first tool calling, MCP workflows·Pop: 80/100

Perf: est. 160 tok/s, typical error 25% · first token ~0.2s

Fits in 20 GB VRAM with room to spare. Best for local ai agents, privacy-first tool calling, mcp workflows on RX 7900 XT.

ollamaregistry-verified
06

Ministral 3 14B

Mistral / 13.5B / Q4_K_M / ~9.1 GB

Local OKExcellentA · 49

Best for: Chat, Reasoning, Vision·Pop: 74/100

Perf: est. 60 tok/s, typical error 10% · first token ~0.6s

Fits in 20 GB VRAM with room to spare. Best for chat, reasoning, vision on RX 7900 XT.

ollamaregistry-verified
07

Devstral Small 2 24B

Mistral / 24B / Q4_K_M / ~15.2 GB

Local OKOKB · 16

Best for: Agentic coding, Software engineering·Pop: 74/100

Perf: est. 39 tok/s, typical error 50% · first token ~1.0s

Fits in 20 GB VRAM with room to spare. Best for agentic coding, software engineering on RX 7900 XT.

ollamaregistry-verified
08

Gemma 4 12B

Gemma / 12B / Q4_K_M / ~8 GB

Local OKExcellentA · 56

Best for: Chat, Coding, Multimodal·Pop: 80/100

Perf: est. 68 tok/s, typical error 10% · first token ~0.6s

Fits in 20 GB VRAM with room to spare. Best for chat, coding, multimodal on RX 7900 XT.

ollamaregistry-verified
09

GPT-OSS 20B

GPT-OSS / 21B / MXFP4 / ~13.8 GB

Local OKOKB · 23

Best for: Chat, Coding, Reasoning·Pop: 85/100

Perf: est. 112 tok/s, typical error 25% · first token ~1.2s

Fits in 20 GB VRAM with room to spare. Best for chat, coding, reasoning on RX 7900 XT.

ollamaregistry-verified
10

Qwen3.5 9B Instruct

Qwen / 9B / Q4_K_M / ~7 GB

Local OKExcellentA · 61

Best for: Quality, Coding, Reasoning·Pop: 86/100

Perf: est. 96 tok/s, typical error 10% · first token ~0.4s

Fits in 20 GB VRAM with room to spare. Best for quality, coding, reasoning on RX 7900 XT.

ollamaregistry-verified

Video and image models on the RX 7900 XT

No stated VRAM figure or published run covers AMD for these models yet. By memory size: video models · image models.

Config reports

No speed test run for this GPU yet.

Share your config

Every report is reviewed before it appears. No email or account needed.

0/280

Compatible model families

RX 7900 XT FAQ: common questions

What size LLM can I run on an RX 7900 XT?

Up to 27B parameter models at Q4 quantization. Its 20GB VRAM gives about 18GB usable, which holds a 27B Q4 (~18GB) fully in VRAM with context kept modest. 32B models need CPU offloading, which cuts speed sharply for dense models.

Does Ollama support the RX 7900 XT?

Yes. RDNA 3 (gfx1100) is officially supported by ROCm, which Ollama uses on Linux and Windows. llama.cpp and LM Studio also work via ROCm/HIP or Vulkan.

RX 7900 XT vs RTX 5070 Ti for local AI?

The 7900 XT has 4GB more VRAM (20 vs 16GB), so it runs 27B-class models the 5070 Ti cannot fit. The 5070 Ti has higher bandwidth (896 vs 800 GB/s) and the CUDA ecosystem, so it is faster on the 14B class both can hold. Pick by target model size.

RX 7900 XT or RX 7900 XTX for local LLMs?

The XTX if the budget allows: 24GB vs 20GB moves the ceiling from 27B to 32B-class models, and its 960 GB/s is 20% faster. The XT wins on value when the models you want stay at or under 27B.

How fast is a 27B-class model on the RX 7900 XT?

The RX 7900 XT's 20GB of VRAM cannot fit a 32B model comfortably. The largest size class it fits is 20B MoE (3.6B active), at est. 114 tok/s, typical error 25%. The largest individual model in the catalog that fits is Muse Glimmer 30B (30B).

Related guides & benchmarks

More local AI guides

Browse all NVIDIA GPUs for AI

Real-world calibration

Have this hardware? Share a measured run.

Time a model that fits your machine with Ollama, using the same prompts as every other run. The CLI shows your result and asks before it sends anything; your machine then joins the speed test leaderboard. It does not change the quality benchmark.

Want personalized recommendations?

Use our interactive wizard to compare models across Apple Silicon and NVIDIA GPUs.