Radeon AI PRO R9700 32G — what LLMs can it run?
125 of 151 indexed models fit comfortably in 32GB at 4K context, each at the highest-quality quant that still leaves headroom.
The short answer for Radeon AI PRO R9700 32G
32 GB, 640 GB/s GDDR6. 125 of 151 models in this index fit comfortably at 4K context; 126 load at all.
- Biggest that fits
- Kolibri-1 78B-A3.5B — 25.6 GB at IQ2_XS — 6.4 GB spare at 4K, so longer context comes out of a thin margin.Download: bartowski/Kolibri-1-GGUF · Kolibri-1-IQ2_XS.gguf · 24.34 GB · listing read 2026-11-12
- Room to grow
- Gemma 3 27B IT — 18.3 GB at Q4_K_M — under 60% of the card, which leaves 13.7 GB for a long context window or a second process.Download: bartowski/google_gemma-3-27b-it-GGUF · google_gemma-3-27b-it-Q4_K_M.gguf · 16.55 GB · + vision projector 0.86 GB, separate · listing read 2026-11-03
- Speed ceiling
- Llama 3.1 8B Instruct at Q4_K_M reads 4.6 GB of weights per token, so 640 GB/s puts a hard ceiling near 139 tok/s. That is arithmetic on two published numbers, not a benchmark — real throughput lands below it. No run on this card has been measured here, so there is nothing to compare the ceiling against.
- The wall you will hit
- Backend coverage, not capacity. At 640 GB/s the memory system is competitive with NVIDIA cards of the same size; what varies is whether your runtime has a working ROCm build for your kernel and card. llama.cpp and vLLM both ship official ROCm support — this is a setup question, not a hardware ceiling.
50B+ · 4
| Model | Quant | Est. VRAM | Headroom | tok/s |
|---|---|---|---|---|
| Kolibri-1 78B-A3.5B78B-A3.5BPair fit details | IQ2_XS | 25.59 GB | +6.4 GB | ≤618Ceiling |
| Nemotron 3.5 Bolt 74B-A9B74B-A9BPair fit details | IQ2_XXS | 23.05 GB | +9 GB | ≤252Ceiling |
| Ripple-65B-Preview65BPair fit details | IQ2_S | 27.18 GB | +4.8 GB | ≤27Ceiling |
| Jamba 1.5 Mini52B-A12B | AWQ INT4 | 27.24 GB | +4.8 GB | ≤112Ceiling |
20–50B · 43
| Model | Quant | Est. VRAM | Headroom | tok/s |
|---|---|---|---|---|
| Mixtral 8x7B Instruct47B MoE | AWQ INT4 | 24.95 GB | +7.1 GB | ≤104Ceiling |
| Recursive Flint 40B-A2.5B40B-A2.5BPair fit details | Q4_K_M | 26.68 GB | +5.3 GB | ≤424Ceiling |
| K2 Horizon MoVA 36B-A4B36B-A4BPair fit details | Q4_K_M | 24.21 GB | +7.8 GB | ≤282Ceiling |
| Seed-OSS 36B Instruct36BPair fit details | Q5_K_M | 27.83 GB | +4.2 GB | ≤26Ceiling |
| Qwen3.6 35B-A3B35B-A3BPair fit details | Q4_K_M | 23.38 GB | +8.6 GB | ≤358Ceiling |
| Apex 35B V1.035B-A3BPair fit details | Q4_K_M | 23.91 GB | +8.1 GB | ≤350Ceiling |
| Command R 35B35B | Q4_K_M | 22.86 GB | +9.1 GB | ≤32Ceiling |
| K2 Horizon 32B32BPair fit details | Q5_K_M | 26.91 GB | +5.1 GB | ≤27Ceiling |
| MiniCPM-V 4.7 35B-A3B35B-A3BPair fit details | Q5_K_M | 25.91 GB | +6.1 GB | ≤315Ceiling |
| Clover-35B-A3B-Preview35B-A3BPair fit details | Q4_K_M | 23.4 GB | +8.6 GB | ≤349Ceiling |
| Yi 1.5 34B Chat34B | Q4_K_M | 22.82 GB | +9.2 GB | ≤32Ceiling |
| Nanbeige4.3 33B33BPair fit details | Q4_K_M | 22.61 GB | +9.4 GB | ≤32Ceiling |
| Qwen3 32B Instruct32B | Q4_K_M | 21.85 GB | +10.1 GB | 18–24Derived |
| Qwen2.5 32B Instruct32B | Q4_K_M | 21.69 GB | +10.3 GB | 18–24Derived |
| Qwen2.5-Coder 32B Instruct32B | Q4_K_M | 21.69 GB | +10.3 GB | 18–24Derived |
| DeepSeek-R1-Distill-Qwen-32B32B | Q4_K_M | 21.69 GB | +10.3 GB | 18–24Derived |
| Nemotron 3.5 Lightning 30B-A3B30B-A3BPair fit details | Q4_0 | 19.79 GB | +12.2 GB | ≤375Ceiling |
| Xing4.0 29B-A4B29B-A4BPair fit details | Q5_K_M | 25.28 GB | +6.7 GB | ≤219Ceiling |
| Gemma 4 31B IT31BPair fit details | Q5_K_M | 25.25 GB | +6.8 GB | 16–21Derived |
| Qwen3-Coder 30B-A3B Instruct30B-A3BPair fit details | Q5_K_M | 23.12 GB | +8.9 GB | ≤316Ceiling |
| Qwen3-VL 30B-A3B Instruct30B-A3BPair fit details | Q4_K_M | 19.89 GB | +12.1 GB | ≤368Ceiling |
| Qwen3 30B-A3B Instruct30B-A3B | Q5_K_M | 23.04 GB | +9 GB | ≤316Ceiling |
| North Mini Code 1.0 30B-A3B30B-A3BPair fit details | Q4_K_M | 20.01 GB | +12 GB | ≤365Ceiling |
| Hy-Memory-30B-A3B30B-A3BPair fit details | Q6_K | 26.21 GB | +5.8 GB | ≤273Ceiling |
| GLM-4.7-Flash30B-A3BPair fit details | Q4_K_M | 19.51 GB | +12.5 GB | ≤364Ceiling |
| Granite 4.2 30B30BPair fit details | Q6_K | 26.19 GB | +5.8 GB | 15–20Derived |
| Muse Glimmer 30B30BPair fit details | Q6_K | 24.61 GB | +7.4 GB | 15–20Derived |
| Qwen3.8 27B27BPair fit details | Q6_K | 25.2 GB | +6.8 GB | 15–20Derived |
| Arete 27B27BPair fit details | Q5_K_M | 21.89 GB | +10.1 GB | 17–23Derived |
| Qwen3.6 27B27BPair fit details | Q5_K_M | 22.19 GB | +9.8 GB | 17–22Derived |
| DaisyRose 27B27BPair fit details | Q5_K_M | 22 GB | +10 GB | 17–23Derived |
| Gemma 2 27B Instruct27B | Q5_K_M | 21.76 GB | +10.2 GB | 18–24Derived |
| Gemma 3 27B IT27BPair fit details | Q4_K_M | 18.3 GB | +13.7 GB | 21–28Derived |
| Ternary Bonsai 2 27B27BPair fit details | Q2_0 | 8.24 GB | +23.8 GB | 47–62Derived |
| Clairic 27B27BPair fit details | Q5_K_M | 20.51 GB | +11.5 GB | 18–24Derived |
| Gemma 4 26B-A4B IT26B-A4BPair fit details | Q4_K_M | 18.22 GB | +13.8 GB | ≤250Ceiling |
| Mistral Small 24B Instruct24B | Q4_K_M | 15.89 GB | +16.1 GB | 24–32Derived |
| LFM2.5 24B-A2B24B-A2BPair fit details | Q8_0 | 26.73 GB | +5.3 GB | ≤259Ceiling |
| Devstral Small 1.1 24B24BPair fit details | Q6_K | 20.89 GB | +11.1 GB | 18–24Derived |
| Magistral Small 1.2 24B24BPair fit details | Q6_K | 20.89 GB | +11.1 GB | 18–24Derived |
| Codestral 22B22B | Q4_K_M | 15.03 GB | +17 GB | 26–35Derived |
| ERNIE 4.5 21B-A3B21B-A3BPair fit details | Q8_0 | 24.48 GB | +7.5 GB | ≤211Ceiling |
| GPT-OSS 20B21B MoE | MXFP4 | 12.76 GB | +19.2 GB | ≤323Ceiling |
10–20B · 19
| Model | Quant | Est. VRAM | Headroom | tok/s |
|---|---|---|---|---|
| InternLM2 20B Chat20B | Q5_K_M | 15.59 GB | +16.4 GB | 25–33Derived |
| DeepSeek-Coder-V2-Lite Instruct16B-A2.4B | Q8_0 | 17.56 GB | +14.4 GB | ≤264Ceiling |
| DeepSeek-V2-Lite Chat16B-A2.4B | Q4_K_M | 10.08 GB | +21.9 GB | ≤463Ceiling |
| StarCoder2 15B15B | Q4_K_M | 9.85 GB | +22.2 GB | 39–52Derived |
| Qwen3 14B Instruct14B | Q5_K_M | 11.65 GB | +20.4 GB | 34–45Derived |
| Qwen2.5 14B Instruct14B | Q5_K_M | 11.73 GB | +20.3 GB | 34–45Derived |
| DeepSeek-R1-Distill-Qwen-14B14B | Q4_K_M | 10.14 GB | +21.9 GB | 40–53Derived |
| Phi-4 14B14B | Q5_K_M | 11.77 GB | +20.2 GB | 34–45Derived |
| Phi-3 Medium 14B Instruct14B | Q6_K | 12.86 GB | +19.1 GB | 31–41Derived |
| Ministral 3 14B Instruct14BPair fit details | Q8_0 | 15.7 GB | +16.3 GB | 25–33Derived |
| Mistral Nemo 12B Instruct12B | Q6_K | 11.14 GB | +20.9 GB | 36–47Derived |
| Gemma 3 12B IT12B | Q5_K_M | 9.84 GB | +22.2 GB | 41–54Derived |
| Mellum2.1 12B-A2.5B Thinking12B-A2.5B | Q8_0 | 13.6 GB | +18.4 GB | ≤254Ceiling |
| Stable LM 2 12B Chat12B | Q4_K_M | 8.52 GB | +23.5 GB | 48–64Derived |
| Gemma 4 12B IT12BPair fit details | Q8_0 | 13.82 GB | +18.2 GB | 28–37Derived |
| Llama 3.2 11B Vision Instruct11B | Q8_0 | 12.9 GB | +19.1 GB | 30–40Derived |
| Solar 10.7B Instruct11B | Q4_K_M | 7.6 GB | +24.4 GB | 55–73Derived |
| Falcon 3 10B Instruct10B | Q4_K_M | 7.28 GB | +24.7 GB | 56–75Derived |
| Metis-1.6 Logos10B MoEPair fit details | Q8_0 | 11.67 GB | +20.3 GB | ≤242Ceiling |
4–10B · 37
| Model | Quant | Est. VRAM | Headroom | tok/s |
|---|---|---|---|---|
| Gemma 2 9B Instruct9B | Q8_0 | 11.7 GB | +20.3 GB | 36–48Derived |
| Qwen3.5 9B9BPair fit details | Q8_0 | 10.39 GB | +21.6 GB | 36–48Derived |
| K2 Horizon 7B7BPair fit details | Q8_0 | 10.62 GB | +21.4 GB | 37–49Derived |
| GLM-4-9B-Chat9B | Q8_0 | 10.16 GB | +21.8 GB | 37–49Derived |
| MiMo-V2.6-Distill-Qwen-9B9BPair fit details | Q8_0 | 10.11 GB | +21.9 GB | 37–49Derived |
| Clairic Flash 9B9BPair fit details | Q8_0 | 10.1 GB | +21.9 GB | 37–49Derived |
| Qwen3-VL 8B Instruct8B | Q8_0 | 10.39 GB | +21.6 GB | 38–50Derived |
| TwIL 8B balanced8.8B | Q4_K_M | 6.28 GB | +25.7 GB | 66–88Derived |
| Granite 4.2 8B8BPair fit details | Q8_0 | 10.46 GB | +21.5 GB | 38–50Derived |
| Ministral 3 8B Instruct8B | Q8_0 | 10.02 GB | +22 GB | 39–52Derived |
| LFM2.5 8B-A1B8B-A1.5BPair fit details | Q8_0 | 9.46 GB | +22.5 GB | ≤422Ceiling |
| Qwen2-VL 7B Instruct7B | Q4_K_M | 5.49 GB | +26.5 GB | 71–94Derived |
| Granite 3.1 8B Instruct8B | Q4_K_M | 5.88 GB | +26.1 GB | 72–95Derived |
| Shot Reader 8B8BPair fit details | Q8_0 | 9.72 GB | +22.3 GB | 41–54Derived |
| Qwen3 8B Instruct8B | Q6_K | 7.64 GB | +24.4 GB | 53–70Derived |
| Llama 3.1 8B Instruct8B | Q8_0 | 9.47 GB | +22.5 GB | 42–55Derived |
| Nous Hermes 3 Llama 3.1 8B8B | Q4_K_M | 5.64 GB | +26.4 GB | 73–97Derived |
| Aya 23 8B8B | Q4_K_M | 5.64 GB | +26.4 GB | 73–97Derived |
| OpenChat 3.6 8B8B | Q4_K_M | 5.64 GB | +26.4 GB | 73–97Derived |
| DeepSeek-R1-Distill-Llama-8B8B | Q5_K_M | 6.51 GB | +25.5 GB | 62–82Derived |
| InternLM2 7B Chat7B | Q4_K_M | 5.45 GB | +26.6 GB | 76–100Derived |
| Qwen2.5 7B Instruct7B | Q6_K | 6.77 GB | +25.2 GB | 57–75Derived |
| Qwen2.5-Coder 7B Instruct7B | Q4_K_M | 5.07 GB | +26.9 GB | 77–102Derived |
| DeepSeek-R1-Distill-Qwen-7B7B | Q4_K_M | 5.07 GB | +26.9 GB | 77–102Derived |
| Gemma 4 E4B ITE4BPair fit details | Q8_0 | 8.68 GB | +23.3 GB | 43–57Derived |
| Nemotron 3.5 Bolt E4BE4BPair fit details | Q8_0 | 8.5 GB | +23.5 GB | 44–59Derived |
| OLMo 2 7B Instruct7B | Q8_0 | 10.3 GB | +21.7 GB | 46–61Derived |
| OLMo 3 7B Instruct7B | Q8_0 | 10.3 GB | +21.7 GB | 46–61Derived |
| WizardLM-2 7B7B | Q4_K_M | 5.14 GB | +26.9 GB | 81–107Derived |
| Mistral 7B Instruct v0.37B | Q6_K | 6.75 GB | +25.3 GB | 60–79Derived |
| Zephyr 7B Beta7B | Q6_K | 6.75 GB | +25.3 GB | 60–79Derived |
| K2 Horizon 3.7B3.7B | Q8_0 | 6.24 GB | +25.8 GB | ≤125Ceiling |
| Gemma 4 E2B ITE2B | Q8_0 | 5.2 GB | +26.8 GB | ≤136Ceiling |
| Qwen3.5 4B4B | Q8_0 | 4.97 GB | +27 GB | ≤146Ceiling |
| Gemma 3 4B IT4B | Q8_0 | 5.05 GB | +27 GB | ≤147Ceiling |
| Spark-X2.6 4B4B | Q8_0 | 4.84 GB | +27.2 GB | ≤154Ceiling |
| Qwen3 4B Instruct4B | Q6_K | 4.05 GB | +28 GB | ≤205Ceiling |
<4B · 22
| Model | Quant | Est. VRAM | Headroom | tok/s |
|---|---|---|---|---|
| Phi-4 Mini Instruct3.8B | Q8_0 | 4.81 GB | +27.2 GB | ≤165Ceiling |
| Phi-3.5 Mini Instruct3.8B | Q8_0 | 5.89 GB | +26.1 GB | ≤166Ceiling |
| Granite 4.2 3B3B | Q8_0 | 4.41 GB | +27.6 GB | ≤173Ceiling |
| Ministral 3 3B Instruct3B | Q8_0 | 4.26 GB | +27.7 GB | ≤184Ceiling |
| Llama 3.2 3B Instruct3B | Q8_0 | 4.05 GB | +28 GB | ≤198Ceiling |
| Qwen2.5 3B Instruct3B | Q8_0 | 3.59 GB | +28.4 GB | ≤205Ceiling |
| SmolLM3 3B3B | Q8_0 | 3.73 GB | +28.3 GB | ≤206Ceiling |
| Metis-1.6 Praxis3B MoE | Q8_0 | 3.4 GB | +28.6 GB | ≤566Ceiling |
| Gemma 2 2B Instruct2B | Q8_0 | 3.34 GB | +28.7 GB | ≤243Ceiling |
| Qwen3.5 2B2B | Q8_0 | 2.23 GB | +29.8 GB | ≤323Ceiling |
| LFM3 1.7B1.7B | Q8_0 | 1.99 GB | +30 GB | ≤366Ceiling |
| LFM3-VL 2B2B | Q8_0 | 1.99 GB | +30 GB | ≤366Ceiling |
| Qwen3 1.7B Instruct1.7B | Q8_0 | 2.39 GB | +29.6 GB | ≤368Ceiling |
| Qwen2.5 1.5B Instruct1.5B | Q8_0 | 1.83 GB | +30.2 GB | ≤413Ceiling |
| Llama 3.2 1B Instruct1B | Q8_0 | 1.51 GB | +30.5 GB | ≤512Ceiling |
| LFM2.5 1.2B Instruct1.2B | Q8_0 | 1.35 GB | +30.7 GB | ≤542Ceiling |
| K2 Horizon 0.9B0.9B | Q8_0 | 1.44 GB | +30.6 GB | ≤587Ceiling |
| Qwen3.5 0.8B0.8B | Q8_0 | 0.94 GB | +31.1 GB | ≤790Ceiling |
| BotPsy 0.8B0.75B | Q8_0 | 0.9 GB | +31.1 GB | ≤831Ceiling |
| Qwen2.5 0.5B Instruct0.5B | Q8_0 | 0.6 GB | +31.4 GB | ≤1306Ceiling |
| MNN-Pocket 0.3B Instruct0.36B | Q8_0 | 0.46 GB | +31.5 GB | ≤1730Ceiling |
| MNN-Pocket 0.3B QAT0.30B | Q4_1 | 0.26 GB | +31.7 GB | ≤3556Ceiling |
How this list is built
Each row is the lowest-perplexity-loss quant of that model whose estimated total — weights plus KV cache at 4K context plus activation buffer — uses at most 88% of the card. That is the calculator's "green" threshold, so each estimate leaves headroom. Raise the context length and the list shortens; the calculator lets you check any combination directly. Speed figures are labelled as measurements, estimates scaled from this card's Llama 2 7B benchmark, or theoretical bandwidth ceilings. A figure that only exists for the RTX 4090 is labelled as that card.
A further 1 models load but with no headroom to spare (up to 105% of VRAM) — the Quant Hub’s GPU chips count those too, which is why its number is higher.
Measured on this card
No benchmark runs in this index were recorded on a Radeon AI PRO R9700 32G. Every figure on this page is calculated from the model architecture and the quant level — treat them as estimates, not measurements.
Cards with the same budget
What fits is decided by memory, so every 32GB card of this type returns the same list. These pages are not different answers — they differ in throughput, which this index does not measure per card.
The same 32GB in system memory is a different machine — everything on this page fits there, at a speed set by your DIMMs rather than a graphics bus: 32 GB RAM (CPU)
Same list, different speed — play one sentence on both cards: Radeon AI PRO R9700 32G vs Instinct MI100 32G (Qwen3 32B Instruct)
Setup guides for this card
Step-by-step guides written for the Radeon AI PRO R9700 32G's runtime or memory size, with the commands and the checks that prove the model is running on the card.
Stepping up
A Radeon PRO W7900 48G (48GB) fits 6 more of the indexed models than this card. Radeon PRO W7900 48G →
Before buying, hear the speed difference on a model both cards hold: Radeon AI PRO R9700 32G vs Radeon PRO W7900 48G (Qwen3 32B Instruct) →
Common questions
What is the best local LLM for a Radeon AI PRO R9700 32G?
For everyday use, Gemma 3 27B IT at Q4_K_M — about 18.3 GB of the card's 32 GB at 4K context, leaving 13.7 GB for a longer window. If you want the largest thing that will load, that is Kolibri-1 78B-A3.5B at IQ2_XS (25.6 GB, 6.4 GB spare). "Best" here means best fit for the memory budget — this index does not run task benchmarks, so it cannot tell you which model is smarter.
Can a Radeon AI PRO R9700 32G run Llama 3.2 90B Vision Instruct?
No. Its smallest build here, Q3_K_M, needs about 46.2 GB and this card has 32.0 GB usable — short by 14.2 GB before any context beyond 4K. The largest model this card does clear is Kolibri-1 78B-A3.5B.
How many tokens per second does a Radeon AI PRO R9700 32G do on an 8B model at Q4?
This index has no measured run on this card, so it does not publish a figure. What can be stated from specifications: generating a token requires reading every weight once, Llama 3.1 8B Instruct at Q4_K_M is 4.6 GB of weights, and this card moves 640 GB/s — a ceiling near 139 tok/s. Batch size, context length, the runtime and how much of the model sits in cache all take you below it.
Radeon AI PRO R9700 32G or RTX 5090 for local LLMs?
They hold the same models: 32 GB against 32 GB fits 125 and 125 of 151 respectively at 4K. The difference is throughput — 640 GB/s against 1,792 GB/s, a 2.8× gap in how fast the weights can be read, which is what token generation is bound by. The RTX 5090 generates faster on any model both can hold.
125 of 151 indexed models fit comfortably in 32GB at 4K context, each at the highest-quality quant that still leaves headroom.
This page's figures change when the model or the runtime does.Last updated 2026-11-22 RSS → /feed.xml