AMD Radeon RX 9070
16 GB · 640 GB/s · 220 W
In short
The AMD Radeon RX 9070 has 16 GB of GDDR6 memory with 640 GB/s of bandwidth, which sets how fast it generates text. At 32k context it runs 61 of the 125 open models we track fully in memory; the most capable is Qwen 3.8 27B at Q6_K, at about 14 tok/s (estimated). We have no current price for the AMD Radeon RX 9070 in the US.
Updated . Speeds are estimates; how they are worked out.
Prices
Every retailer we track in the US, cheapest first. Buy links may earn us a commission.
No retailer in this market lists it right now.
What two of them run
Every model in the catalogue, with the best quant for this hardware. Estimates, not measurements.
| Model | Quant | Fits | Memory | Max context | Decode @32k | Prefill | Good for |
|---|---|---|---|---|---|---|---|
| Qwen 3.8 27B Alibaba Qwen | Q6_K | fits | 25.9 GB | 98k172k with q8 cache | 14tok/sest. | 204 tok/s | – |
| Qwen 3.6 27B Alibaba Qwen | Q8_0 | fits | 30.3 GB | 34k59k with q8 cache | 11tok/sest. | 204 tok/s | – |
| Gemma 4 31B | Q6_K | fits | 28.2 GB | 57k110k with q8 cache | 12tok/sest. | 182 tok/s | – |
| Qwen 3.6 35B-A3B Alibaba Qwen | UD-Q6_K | fits | 29.5 GB | 68k107k with q8 cache | 66tok/sest. | 1455 tok/s | chat, coding, agents, long documents |
| Qwen 3.5 27B Alibaba Qwen | Q8_0 | fits | 30.3 GB | 34k59k with q8 cache | 11tok/sest. | 204 tok/s | – |
| Gemma 4 26B-A4B | Q8_0 | fits | 27.3 GB | 153k245k with q8 cache | 42tok/sest. | 1294 tok/s | chat, coding, agents, long documents |
| Qwen 3.5 35B-A3B Alibaba Qwen | Q6_K | fits | 29.0 GB | 85k133k with q8 cache | 67tok/sest. | 1455 tok/s | chat, coding, agents, long documents |
| GLM 4.7 Flash Z.ai | Q6_K | fits | 26.1 GB | 106k183k with q8 cache | 36tok/sest. | 760 tok/s | chat, coding, agents, long documents |
| Devstral Small 2 24B Mistral AI | Q8_0 | fits | 29.8 GB | 35k64k with q8 cache | 12tok/sest. | 215 tok/s | – |
| Nemotron 3.5 Lightning 30B-A3B NVIDIA | UD-Q5_K_M | fits | 29.9 GB | 78k101k with q8 cache | 68tok/sest. | 1660 tok/s | chat, coding, agents, long documents |
| Kimi Linear 48B-A3B Moonshot AI | Q4_K_M | fits | 29.8 GB | 78k107k with q8 cache | 99tok/sest. | 1254 tok/s | chat, coding, long documents |
| gpt-oss 20B OpenAI | MXFP4 | fits | 13.5 GB | 128k | 65tok/sest. | 1086 tok/s | chat, coding, agents, long documents |
| Qwen3 Coder 30B-A3B Alibaba Qwen | Q6_K | fits | 27.9 GB | 57k102k with q8 cache | 37tok/sest. | 669 tok/s | chat, coding, agents |
| Qwen3 32B Alibaba Qwen | Q4_K_M | fits | 27.9 GB | 40k | 12tok/sest. | 127 tok/s | chat |
| Gemma 4 12B | Q8_0 | fits | 14.1 GB | 256k | 26tok/sest. | 473 tok/s | chat, coding, long documents |
| Qwen 3.5 9B Alibaba Qwen | Q8_0 | fits | 11.4 GB | 256k | 33tok/sest. | 589 tok/s | chat, coding, long documents |
| Gemma 3 27B | Q6_K | fits | 25.0 GB | 95k128k with q8 cache | 14tok/sest. | 216 tok/s | chat |
| DeepSeek R1 Distill Qwen 32B DeepSeek | Q4_K_M | fits | 28.0 GB | 41k76k with q8 cache | 12tok/sest. | 145 tok/s | chat |
| Mistral Small 3.2 24B Mistral AI | Q8_0 | fits | 29.8 GB | 35k64k with q8 cache | 12tok/sest. | 215 tok/s | – |
| Ministral 3 14B Mistral AI | Q8_0 | fits | 19.9 GB | 97k176k with q8 cache | 18tok/sest. | 327 tok/s | chat |
| Qwen3 14B Alibaba Qwen | Q8_0 | fits | 21.1 GB | 40k | 17tok/sest. | 294 tok/s | chat |
| DeepSeek R1 0528 Qwen3 8B DeepSeek | Q8_0 | fits | 14.1 GB | 128k | 26tok/sest. | 485 tok/s | chat, coding |
| Qwen3 8B Alibaba Qwen | Q8_0 | fits | 14.1 GB | 40k | 26tok/sest. | 485 tok/s | chat, coding |
| Phi-4 14B Microsoft | Q8_0 | fits | 22.2 GB | 16k | 16tok/sest. | 296 tok/s | chat |
| Qwen 3.5 4B Alibaba Qwen | Q8_0 | fits | 6.7 GB | 256k | 63tok/sest. | 1102 tok/s | chat, coding, long documents |
| Ministral 3 8B Mistral AI | Q8_0 | fits | 14.1 GB | 149k256k with q8 cache | 26tok/sest. | 469 tok/s | chat, coding |
| Granite 4.2 8B IBM | Q8_0 | fits | 15.2 GB | 125k128k with q8 cache | 24tok/sest. | 446 tok/s | chat, coding |
| Gemma 4 E4B | Q8_0 | fits | 10.0 GB | 128k | 38tok/sest. | 740 tok/s | chat, coding, long documents |
| LFM2.5 8B-A1B Liquid AI | Q8_0 | fits | 11.4 GB | 125k | 79tok/sest. | 2420 tok/s | chat, coding, long documents |
| Llama 3.1 8B Meta | Q6_K | fits | 11.6 GB | 128k | 32tok/sest. | 513 tok/s | chat, coding |
| Qwen2.5 0.5B Alibaba Qwen | Q8_0 | fits | 2.4 GB | 32k | 377tok/sest. | 5270 tok/s | chat, coding, long documents |
| Qwen3.5 0.8B Alibaba Qwen | Q8_0 | fits | 2.6 GB | 256k | 285tok/sest. | 4951 tok/s | chat, coding, agents, long documents |
| Qwen3.5 2B Alibaba Qwen | Q5_K_M | fits | 3.2 GB | 256k | 190tok/sest. | 2360 tok/s | chat, coding, agents, long documents |
| DeepSeek R1 Distill Qwen 1.5B DeepSeek | Q8_0 | fits | 4.1 GB | 128k | 124tok/sest. | 1983 tok/s | chat, coding, long documents |
| MiniCPM5 2B openbmb | Q8_0 | fits | 5.3 GB | 128k | 86tok/sest. | 1184 tok/s | chat, coding, agents, long documents |
| Spark X2.5 4B XHToken | mlx-4bit | fits | 5.5 GB | 639k1024k with q8 cache | 82tok/sest. | 1163 tok/s | chat, coding, agents, long documents |
| LFM2.5 2.6B Liquid AI | Q8_0 | fits | 6.0 GB | 128k | 72tok/sest. | 1341 tok/s | chat, coding, agents, long documents |
| LFM2.5 VL 3B Liquid AI | Q8_0 | fits | 6.0 GB | 32k | 72tok/sest. | 1230 tok/s | chat, coding, agents, long documents |
| Mellum2.1 12B A2.5B Thinking JetBrains | Q8_0 | fits | 14.0 GB | 128k | 67tok/sest. | 1747 tok/s | chat, coding, agents, long documents |
| FrogNano 4B 2609 Microsoft | Q8_0 | fits | 6.8 GB | 256k | 61tok/sest. | 1102 tok/s | chat, coding, agents, long documents |
| granite vision 4.1 4b IBM | Q8_0 | fits | 7.4 GB | 128k | 56tok/sest. | 859 tok/s | chat, coding, long documents |
| granite 4.2 3b IBM | Q8_0 | fits | 7.6 GB | 128k | 54tok/sest. | 901 tok/s | chat, coding, agents, long documents |
| Kimi VL A3B Thinking 2506 Moonshot AI | Q8_0 | fits | 18.2 GB | 128k | 51tok/sest. | 1313 tok/s | chat, coding, long documents |
| Kimi VL A3B Moonshot AI | mlx-8bit | fits | 19.0 GB | 128k | 49tok/sest. | 1313 tok/s | chat, coding, agents, long documents |
| Ministral 3 3B Instruct 2512 Mistral AI | Q8_0 | fits | 8.1 GB | 239k256k with q8 cache | 49tok/sest. | 861 tok/s | chat, coding, agents, long documents |
| LFM2.5 24B A2B Liquid AI | Q8_0 | fits | 27.7 GB | 63k113k with q8 cache | 44tok/sest. | 1349 tok/s | chat, coding, agents |
| GLM 4 9B 0414 Z.ai | mlx-5bit | fits | 10.4 GB | 32k | 37tok/sest. | 428 tok/s | chat, coding |
| DeepSeek R1 Distill Qwen 7B DeepSeek | Q8_0 | fits | 10.8 GB | 128k | 35tok/sest. | 579 tok/s | chat, coding, long documents |
| GLM 4.6V Flash Z.ai | Q8_0 | fits | 12.1 GB | 128k | 31tok/sest. | 403 tok/s | chat, coding |
| DeepSeek R1 Distill Llama 8B DeepSeek | Q8_0 | fits | 13.4 GB | 128k | 27tok/sest. | 513 tok/s | chat, coding |
| GLM 4.1V 9B Thinking Z.ai | mlx-8bit | fits | 13.7 GB | 64k | 27tok/sest. | 403 tok/s | chat, coding |
| Ministral 3 8B Reasoning 2512 Mistral AI | Q8_0 | fits | 14.1 GB | 149k256k with q8 cache | 26tok/sest. | 469 tok/s | chat, coding |
| Ministral 3 14B Reasoning 2512 Mistral AI | Q8_0 | fits | 19.9 GB | 97k176k with q8 cache | 18tok/sest. | 327 tok/s | chat |
| DeepSeek R1 Distill Qwen 14B DeepSeek | Q8_0 | fits | 22.1 GB | 74k128k with q8 cache | 16tok/sest. | 277 tok/s | chat |
| Qwen3.8 27B OBLITERATED OBLITERATUS | mlx-4bit | fits | 22.9 GB | 142k248k with q8 cache | 15tok/sest. | 204 tok/s | chat |
| Muse Glimmer 30B meta-models | Q6_K | fits | 23.8 GB | 128k | 15tok/sest. | 196 tok/s | – |
| granite 4.2 30b IBM | Q5_K_M | fits | 29.1 GB | 37k68k with q8 cache | 12tok/sest. | 167 tok/s | chat |
| Devstral Small 2507 Mistral AI | Q8_0 | fits | 29.8 GB | 35k64k with q8 cache | 12tok/sest. | 218 tok/s | – |
| Magistral Small 2506 Mistral AI | Q8_0 | fits | 29.8 GB | 35k40k with q8 cache | 12tok/sest. | 218 tok/s | – |
| Magistral Small 2507 Mistral AI | Q8_0 | fits | 29.8 GB | 35k40k with q8 cache | 12tok/sest. | 218 tok/s | – |
| Magistral Small 2509 Mistral AI | Q8_0 | fits | 29.8 GB | 35k64k with q8 cache | 12tok/sest. | 215 tok/s | – |
| Qwen 3.5 122B-A10B Alibaba Qwen | UD-Q2_K_XL | offload | 41.3 GB | – | 20tok/sest. | 399 tok/s | chat |
| Qwen3 Coder Next Alibaba Qwen | Q4_K_M | offload | 47.5 GB | – | 21tok/sest. | 895 tok/s | chat, coding |
| Nemotron 3 Super 120B-A12B NVIDIA | UD-Q2_K_XL | offload | 52.8 GB | – | 11tok/sest. | 357 tok/s | – |
| Mistral Small 4 119B Mistral AI | UD-Q2_K_XL | offload | 39.6 GB | – | 29tok/sest. | 454 tok/s | chat, coding |
| GLM 4.5 Air Z.ai | Q2_K | offload | 49.4 GB | – | 6tok/sest. | 150 tok/s | – |
| Qwen3 Next 80B-A3B Alibaba Qwen | Q4_K_M | offload | 47.5 GB | – | 26tok/sest. | 1065 tok/s | chat, coding |
| Llama 3.3 70B Meta | Q4_K_M | offload | 51.1 GB | – | 2tok/sest. | 52 tok/s | – |
| DeepSeek R1 Distill Llama 70B DeepSeek | Q4_K_M | offload | 51.1 GB | – | 2tok/sest. | 52 tok/s | – |
| DeepSeek V4 Flash DeepSeek | UD-Q2_K_XL | won't run | 94.4 GB | – | – | – | – |
| GLM 5.3 Flash Z.ai | UD-Q2_K_XL | won't run | 103.2 GB | – | – | – | – |
| Qwen 3.8 Flash Next Alibaba Qwen | UD-Q2_K_XL | won't run | 75.8 GB | – | – | – | – |
| MiniMax M2.7 MiniMax | UD-Q2_K_XL | won't run | 79.4 GB | – | – | – | – |
| gpt-oss 120B OpenAI | Q4_K_M | won't run | 61.1 GB | – | – | – | – |
| DeepSeek R1 DeepSeek | Q4_K_M | won't run | 380.3 GB | – | – | – | – |
| DeepSeek R1 0528 DeepSeek | Q4_K_M | won't run | 380.8 GB | – | – | – | – |
| DeepSeek R1 Zero DeepSeek | Q4_K_M | won't run | 380.3 GB | – | – | – | – |
| DeepSeek V3 DeepSeek | Q4_K_M | won't run | 380.3 GB | – | – | – | – |
| DeepSeek V3 0324 DeepSeek | Q4_K_M | won't run | 380.3 GB | – | – | – | – |
| DeepSeek V3.1 DeepSeek | Q4_K_M | won't run | 381.2 GB | – | – | – | – |
| DeepSeek V3.1 Terminus DeepSeek | Q4_K_M | won't run | 381.2 GB | – | – | – | – |
| DeepSeek V3.2 DeepSeek | Q4_K_M | won't run | 381.2 GB | – | – | – | – |
| DeepSeek V4 Flash 0731 DeepSeek | MXFP4 | won't run | 149.8 GB | – | – | – | – |
| DeepSeek V4 Flash Vision Exp DeepSeek | MXFP4 | won't run | 158.6 GB | – | – | – | – |
| Devstral 2 123B Instruct 2512 Mistral AI | Q4_K_M | won't run | 82.2 GB | – | – | – | – |
| GLM 4.5 Z.ai | Q4_K_M | won't run | 214.6 GB | – | – | – | – |
| GLM 4.5V Z.ai | Q4_K_M | won't run | 66.5 GB | – | – | – | – |
| GLM 4.6 Z.ai | Q4_K_M | won't run | 213.8 GB | – | – | – | – |
| GLM 4.6V Z.ai | Q4_K_M | won't run | 72.8 GB | – | – | – | – |
| GLM 4.7 Z.ai | Q4_K_M | won't run | 214.6 GB | – | – | – | – |
| GLM 5 Z.ai | MXFP4 | won't run | 386.6 GB | – | – | – | – |
| GLM 5.1 Z.ai | MXFP4 | won't run | 424.2 GB | – | – | – | – |
| GLM 5.2 Z.ai | UD-Q4_K_M | won't run | 438.1 GB | – | – | – | – |
| GLM 5.3 Z.ai | Q8_0 | won't run | 750.6 GB | – | – | – | – |
| Kimi Dev 72B Moonshot AI | Q6_K | won't run | 71.4 GB | – | – | – | – |
| Kimi K2 Moonshot AI | Q4_K_M | won't run | 581.8 GB | – | – | – | – |
| Kimi K2 Instruct 0905 Moonshot AI | Q4_K_M | won't run | 582.2 GB | – | – | – | – |
| Kimi K2 Thinking Moonshot AI | Q4_K_M | won't run | 582.2 GB | – | – | – | – |
| Kimi K2.5 Moonshot AI | Q4_K_M | won't run | 582.2 GB | – | – | – | – |
| MiniMax M2 MiniMax | Q4_K_M | won't run | 138.1 GB | – | – | – | – |
| MiniMax M2.1 MiniMax | Q4_K_M | won't run | 138.1 GB | – | – | – | – |
| MiniMax M2.5 MiniMax | MXFP4 | won't run | 124.5 GB | – | – | – | – |
| MiniMax M3 MiniMax | MXFP4 | won't run | 244.9 GB | – | – | – | – |
| Mistral Medium 3.5 128B Mistral AI | Q4_K_M | won't run | 82.2 GB | – | – | – | – |
Specifications
From the manufacturer's published specifications; sources are linked where we have them.
- GPU
- Radeon RX 9070
- Memory
- 16 GB GDDR6
- Memory bus
- 256-bit
- Memory bandwidth
- 640 GB/s
- FP16 compute
- 36.1 TFLOPS
- Software
- ROCm
- Board power
- 220 W
- PCIe generation
- Gen 5
- PCIe lanes
- x16
Compare with
Questions
- How much memory does the AMD Radeon RX 9070 have for AI models?
- The AMD Radeon RX 9070 has 16 GB of GDDR6 memory with 640 GB/s of bandwidth, which sets how fast it generates text.
- What is the best model the AMD Radeon RX 9070 can run locally?
- Qwen 3.8 27B (Q6_K) fits fully in its memory at 32k context and generates at about 14 tok/s, an estimate from its memory bandwidth. 61 of the 125 models we track fit at that context.
- Can the AMD Radeon RX 9070 run DeepSeek V4 Flash?
- No: even at UD-Q2_K_XL, DeepSeek V4 Flash needs about 94 GB at 32k context, more than the AMD Radeon RX 9070 and system RAM can hold.
- How much does the AMD Radeon RX 9070 cost in the US?
- We have no current price for the AMD Radeon RX 9070 in the US.
Speeds on this page are for 32k context. See the best hardware for local AI at every budget.