01
GPT-OSS 20B
GPT-OSS / 21B / MXFP4 / ~13.8 GB
Best for: Chat, Coding, Reasoning·Pop: 85/100
Perf: ~62 tok/s · first token ~0.4s
Local OKOKB · 23
Fits in 20 GB VRAM with room to spare. Best for chat, coding, reasoning on RX 7900 XT.
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.
Quick answer
The best local LLM for the RX 7900 XT is GPT-OSS 20B at ~62 tok/s on its 20GB VRAM. It uses ~13.8GB of VRAM; the RX 7900 XT handles up to 27B parameter models at Q4. A 14B model at Q4 runs at ~46 tok/s. (ModelFit, 2026)
Sizing rule: a Q4 model needs about 0.6 GB of VRAM per billion parameters, and ModelFit budgets 90% of this card's 20GB for weights, context, and KV-cache. The per-size table below uses that same budget. Other strong fits: LFM2 24B-A2B Instruct (24B, ~14GB) and Qwen3.5 9B Instruct (Q8) (9B, ~10.7GB). GPT-OSS 20B runs at an estimated 62 tok/s on this card. 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 gpt-oss:20b
TOP PICK
GPT-OSS 20B
EST. SPEED
~62 tok/s
VRAM NEEDED
~13.8 GB
Speeds are ModelFit estimates from memory bandwidth and model size, not measured benchmarks.
Cite this page: ModelFit, RX 7900 XT 20GB Local LLM: Runs 27B, Top Pick ~62 tok/s, https://modelfit.io/gpu/rx-7900-xt/, updated October 2026, CC BY 4.0.
Last updated: October 3, 2026 · Editor: ModelFit Team
*Used market price
RX 7900 XT Estimated Tokens/sec by Model Size
Q4_K_M · ModelFit estimate| Model Size | Est. Speed | Fit on 20GB |
|---|---|---|
| 7B | ~83 tok/s | Fits in VRAM |
| 14B | ~46 tok/s | Fits in VRAM |
| 20B MoE (3.6B active) | ~70 tok/s | Fits in VRAM |
| 32B | ~5 tok/s | CPU offload (slow) |
| 35B MoE (3B active) | ~45 tok/s | CPU offload (slow) |
| 70B | ~1 tok/s | CPU offload (slow) |
| 120B MoE (5.1B active) | ~10 tok/s | CPU offload (slow) |
Estimated Speed by Model Size
Fits in memoryCPU offload
ModelFit estimates, not measured benchmarks: anchored to an 8B-class Q4_K_M model at 16K context on the RX 7900 XT's 800 GB/s bandwidth, then scaled by model size. 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.
Context costs VRAM too. GPT-OSS 20B loads ~13.8 GB of weights; at 16k context the KV cache adds ~4.0 GB (still fits the ~18 GB usable VRAM), and at 64k it adds ~16.0 GB (exceeds the budget, use a smaller quant or a q8_0 KV cache).
Weights 14 GBKV cache · 16k context 4.0 GBHeadroom 0.2 GBRuntime reserve 2.0 GB
The model leaves about 0.2 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.
Get told when a model beats GPT-OSS 20B on your RX 7900 XT
One email when a new open-weight model is a better fit for 20 GB of VRAM, plus the Thursday weekly.
Where to Buy the RX 7900 XT
≈ $550 street · Used market priceStorage & accessories for your model library
Internal NVMe SSD · 2TB
A Gen4 M.2 drive keeps your whole GGUF and quant collection on fast local storage, loading models straight off NVMe.
Check price on AmazonUSB4 NVMe Enclosure
40Gbps external storage fast enough to run models from. Pair it with an M.2 drive for a portable model vault.
Check price on AmazonModelFit may earn a commission on purchases through these links, at no extra cost to you. Prices shown are approximate street references.
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 by $150-300, making it a strong value pick for the 14B-27B class.
RX 7900 XT vs Top GPUs
| Hardware | Memory | Speed | Bandwidth | Price |
|---|---|---|---|---|
| RX 7900 XT | 20 GB | 74 tok/s | 800 GB/s | $550 |
| RX 7900 XTX | 24 GB | 89 tok/s | 960 GB/s | $700 |
| RTX 3090 | 24 GB | 87 tok/s | 936 GB/s | $900 |
| RTX 5070 Ti | 16 GB | 87 tok/s | 896 GB/s | $1,152 |
Recommended Models
registry-verified10 models02
LFM2 24B-A2B Instruct
LFM2 / 24B / Q4_K_M / ~14 GB
Best for: Local AI agents, privacy-first tool calling, MCP workflows·Pop: 80/100
Perf: ~84 tok/s · first token ~0.4s
Local OKOKB · 22
Fits in 20 GB VRAM with room to spare. Best for local ai agents, privacy-first tool calling, mcp workflows on RX 7900 XT.
03
Qwen3.5 9B Instruct (Q8)
Qwen / 9B / Q8_0 / ~10.7 GB
Best for: Quality, Coding, Reasoning·Pop: 86/100
Perf: ~42 tok/s · first token ~0.4s
Local OKExcellentA · 41
Fits in 20 GB VRAM with room to spare. Best for quality, coding, reasoning on RX 7900 XT.
04
Gemma 4 12B (Q8)
Gemma / 12B / Q8_0 / ~12.8 GB
Best for: Chat, Coding, Multimodal·Pop: 80/100
Perf: ~33 tok/s · first token ~0.5s
Local OKOKB · 29
Fits in 20 GB VRAM with room to spare. Best for chat, coding, multimodal on RX 7900 XT.
05
Qwen3 14B
Qwen / 14B / Q4_K_M / ~11 GB
Best for: Coding, Quality·Pop: 84/100
Perf: ~46 tok/s · first token ~0.4s
Local OKExcellentA · 39
Fits in 20 GB VRAM with room to spare. Best for coding, quality on RX 7900 XT.
06
Gemma 3 12B Instruct
Gemma / 12B / Q4_K_M / ~9.5 GB
Best for: Chat, Quality·Pop: 76/100
Perf: ~52 tok/s · first token ~0.4s
Local OKExcellentA · 47
Fits in 20 GB VRAM with room to spare. Best for chat, quality on RX 7900 XT.
07
Ministral 3 14B
Mistral / 13.5B / Q4_K_M / ~9.1 GB
Best for: Chat, Reasoning, Vision·Pop: 74/100
Perf: ~47 tok/s · first token ~0.4s
Local OKExcellentA · 49
Fits in 20 GB VRAM with room to spare. Best for chat, reasoning, vision on RX 7900 XT.
08
Gemma 4 26B-A4B
Gemma / 26B / Q4_K_M / ~16 GB
Best for: Chat, Coding, Multimodal·Pop: 86/100
Perf: ~60 tok/s · first token ~0.4s
Local OKOKB · 11
Fits in 20 GB VRAM with room to spare. Best for chat, coding, multimodal on RX 7900 XT.
09
Mistral Nemo 12B
Mistral / 12B / Q4_K_M / ~9.5 GB
Best for: Chat, Translation·Pop: 78/100
Perf: ~52 tok/s · first token ~0.4s
Local OKExcellentA · 47
Fits in 20 GB VRAM with room to spare. Best for chat, translation on RX 7900 XT.
10
Devstral Small 2 24B
Mistral / 24B / Q4_K_M / ~15.2 GB
Best for: Agentic coding, Software engineering·Pop: 74/100
Perf: ~29 tok/s · first token ~0.5s
Local OKOKB · 16
Fits in 20 GB VRAM with room to spare. Best for agentic coding, software engineering on RX 7900 XT.
Compare With Top GPUs
Compatible Model Families
Qwen
Alibaba Cloud: Qwen3.8 27B local flagship (Apache 2.0, vision)
LlamaMeta: Most popular open-weight model family
DeepSeekDeepSeek AI: V4.1 Flash: multimodal 552B MoE, API only
MistralMistral AI: Excellent performance-per-parameter ratio
GemmaGoogle DeepMind: Gemma 4 E4B: 4.5B effective, runs on 8GB Macs
PhiMicrosoft: Best quality-per-gigabyte at small sizes
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 an estimated 70 tok/s. The largest individual model in the catalog that fits is Muse Glimmer 30B (30B).
Browse All NVIDIA GPUs for AI
Real-world calibration
Have this hardware? Share a measured run.
Measure a fixed Ollama reference model on your machine. This separate, opt-in contribution does not change ModelFit's third-party SWE-Bench leaderboard; results appear on the measured-speed leaderboard after the next site update.
$ npx -y @wecko-ai/modelfit bench --submit --cleanup
CLI setup and package details →See measured machines →Want Personalized Recommendations?
Use our interactive wizard to compare models across Apple Silicon and NVIDIA GPUs.