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 8k context it runs 37 of the 125 open models we track fully in memory; the most capable is gpt-oss 20B at MXFP4, at about 80 tok/s (estimated). We have no current price for the AMD Radeon RX 9070 in Germany.
Updated . Speeds are estimates; how they are worked out.
Prices
Every retailer we track in Germany, cheapest first. Buy links may earn us a commission.
No retailer in this market lists it right now.
What it runs
Every model in the catalogue, with the best quant for this hardware. Estimates, not measurements.
| Model | Quant | Fits | Memory | Max context | Decode @8k | Prefill | Good for |
|---|---|---|---|---|---|---|---|
| gpt-oss 20B OpenAI | MXFP4 | fits | 12.3 GB | 106k128k with q8 cache | 80tok/sest. | 1371 tok/s | chat, coding, agents, long documents |
| Gemma 4 12B | Q8_0 | fits | 13.1 GB | 106k171k with q8 cache | 27tok/sest. | 503 tok/s | chat, coding, long documents |
| Qwen 3.5 9B Alibaba Qwen | Q8_0 | fits | 10.0 GB | 147k243k with q8 cache | 36tok/sest. | 637 tok/s | chat, coding, long documents |
| Ministral 3 14B Mistral AI | Q6_K | fits | 12.4 GB | 25k45k with q8 cache | 28tok/sest. | 413 tok/s | chat |
| Qwen3 14B Alibaba Qwen | Q6_K | fits | 13.4 GB | 19k35k with q8 cache | 26tok/sest. | 384 tok/s | chat |
| DeepSeek R1 0528 Qwen3 8B DeepSeek | Q8_0 | fits | 10.1 GB | 42k78k with q8 cache | 36tok/sest. | 672 tok/s | chat, coding |
| Qwen3 8B Alibaba Qwen | Q8_0 | fits | 10.1 GB | 40k | 36tok/sest. | 672 tok/s | chat, coding |
| Phi-4 14B Microsoft | Q6_K | fits | 13.6 GB | 15k16k with q8 cache | 26tok/sest. | 387 tok/s | chat |
| Qwen 3.5 4B Alibaba Qwen | Q8_0 | fits | 5.3 GB | 256k | 73tok/sest. | 1282 tok/s | chat, coding, long documents |
| Ministral 3 8B Mistral AI | Q8_0 | fits | 10.3 GB | 43k78k with q8 cache | 35tok/sest. | 628 tok/s | chat, coding |
| Granite 4.2 8B IBM | Q8_0 | fits | 10.8 GB | 35k64k with q8 cache | 33tok/sest. | 623 tok/s | chat, coding |
| Gemma 4 E4B | Q8_0 | fits | 8.7 GB | 128k | 42tok/sest. | 772 tok/s | chat, coding, long documents |
| LFM2.5 8B-A1B Liquid AI | Q8_0 | fits | 9.6 GB | 113k125k with q8 cache | 144tok/sest. | 4504 tok/s | chat, coding, long documents |
| Llama 3.1 8B Meta | Q6_K | fits | 8.0 GB | 63k114k with q8 cache | 46tok/sest. | 694 tok/s | chat, coding |
| Qwen2.5 0.5B Alibaba Qwen | Q8_0 | fits | 1.4 GB | 32k | 557tok/sest. | 9426 tok/s | chat, coding, long documents |
| Qwen3.5 0.8B Alibaba Qwen | Q8_0 | fits | 1.7 GB | 256k | 378tok/sest. | 6486 tok/s | chat, coding, agents, long documents |
| Qwen3.5 2B Alibaba Qwen | Q5_K_M | fits | 2.3 GB | 256k | 226tok/sest. | 2660 tok/s | chat, coding, agents, long documents |
| DeepSeek R1 Distill Qwen 1.5B DeepSeek | Q8_0 | fits | 2.8 GB | 128k | 165tok/sest. | 2967 tok/s | chat, coding, long documents |
| MiniCPM5 2B openbmb | Q8_0 | fits | 3.7 GB | 128k | 116tok/sest. | 1961 tok/s | chat, coding, agents, long documents |
| LFM2.5 2.6B Liquid AI | Q8_0 | fits | 4.0 GB | 128k | 104tok/sest. | 1974 tok/s | chat, coding, agents, long documents |
| LFM2.5 VL 3B Liquid AI | Q8_0 | fits | 4.0 GB | 32k | 104tok/sest. | 1742 tok/s | chat, coding, agents, long documents |
| Spark X2.5 4B XHToken | mlx-4bit | fits | 4.0 GB | 280k470k with q8 cache | 104tok/sest. | 1395 tok/s | chat, coding, agents, long documents |
| Kimi VL A3B Thinking 2506 Moonshot AI | Q4_K_M | fits | 10.9 GB | 128k | 97tok/sest. | 1830 tok/s | chat, coding, long documents |
| granite vision 4.1 4b IBM | Q8_0 | fits | 4.8 GB | 128k | 82tok/sest. | 1306 tok/s | chat, coding, long documents |
| Ministral 3 3B Instruct 2512 Mistral AI | Q8_0 | fits | 5.1 GB | 102k183k with q8 cache | 78tok/sest. | 1338 tok/s | chat, coding, agents |
| granite 4.2 3b IBM | Q8_0 | fits | 5.1 GB | 128k | 77tok/sest. | 1404 tok/s | chat, coding, agents, long documents |
| Kimi VL A3B Moonshot AI | mlx-6bit | fits | 14.0 GB | 42k70k with q8 cache | 76tok/sest. | 1830 tok/s | chat, coding, agents, long documents |
| Mellum2.1 12B A2.5B Thinking JetBrains | Q8_0 | fits | 13.0 GB | 118k128k with q8 cache | 75tok/sest. | 2170 tok/s | chat, coding, agents, long documents |
| FrogNano 4B 2609 Microsoft | Q8_0 | fits | 5.4 GB | 256k | 71tok/sest. | 1282 tok/s | chat, coding, agents, long documents |
| DeepSeek R1 Distill Qwen 7B DeepSeek | Q8_0 | fits | 8.8 GB | 113k128k with q8 cache | 41tok/sest. | 749 tok/s | chat, coding |
| GLM 4 9B 0414 Z.ai | mlx-5bit | fits | 8.8 GB | 32k | 41tok/sest. | 588 tok/s | chat, coding |
| DeepSeek R1 Distill Llama 8B DeepSeek | Q8_0 | fits | 9.8 GB | 49k89k with q8 cache | 37tok/sest. | 694 tok/s | chat, coding |
| Ministral 3 8B Reasoning 2512 Mistral AI | Q8_0 | fits | 10.3 GB | 43k78k with q8 cache | 35tok/sest. | 628 tok/s | chat, coding |
| GLM 4.6V Flash Z.ai | Q8_0 | fits | 10.5 GB | 113k128k with q8 cache | 34tok/sest. | 543 tok/s | chat, coding |
| GLM 4.1V 9B Thinking Z.ai | mlx-8bit | fits | 12.1 GB | 64k | 29tok/sest. | 543 tok/s | chat, coding |
| Ministral 3 14B Reasoning 2512 Mistral AI | Q6_K | fits | 12.4 GB | 25k45k with q8 cache | 28tok/sest. | 413 tok/s | chat |
| DeepSeek R1 Distill Qwen 14B DeepSeek | Q6_K | fits | 13.6 GB | 16k29k with q8 cache | 26tok/sest. | 376 tok/s | chat |
| Qwen 3.8 27B Alibaba Qwen | UD-Q4_K_M | offload | 16.8 GB | – | 13tok/sest. | 208 tok/s | – |
| Qwen 3.5 122B-A10B Alibaba Qwen | UD-Q2_K_XL | offload | 40.1 GB | – | 13tok/sest. | 362 tok/s | – |
| Qwen 3.6 27B Alibaba Qwen | Q4_K_M | offload | 17.1 GB | – | 12tok/sest. | 205 tok/s | – |
| Gemma 4 31B | Q4_K_M | offload | 19.3 GB | – | 8tok/sest. | 167 tok/s | – |
| Mistral Small 4 119B Mistral AI | UD-Q2_K_XL | offload | 38.5 GB | – | 19tok/sest. | 481 tok/s | chat |
| Qwen 3.6 35B-A3B Alibaba Qwen | MXFP4 | offload | 21.3 GB | – | 39tok/sest. | 1556 tok/s | chat, coding |
| Qwen 3.5 27B Alibaba Qwen | Q4_K_M | offload | 17.1 GB | – | 12tok/sest. | 206 tok/s | – |
| Gemma 4 26B-A4B | MXFP4 | offload | 16.6 GB | – | 48tok/sest. | 1366 tok/s | chat, coding, agents, long documents |
| Qwen 3.5 35B-A3B Alibaba Qwen | Q4_K_M | offload | 21.6 GB | – | 38tok/sest. | 1545 tok/s | chat, coding |
| GLM 4.7 Flash Z.ai | Q4_K_M | offload | 18.3 GB | – | 31tok/sest. | 947 tok/s | chat |
| Devstral Small 2 24B Mistral AI | Q4_K_M | offload | 15.4 GB | 6k11k with q8 cache | 21tok/sest. | 247 tok/s | chat |
| Nemotron 3.5 Lightning 30B-A3B NVIDIA | MXFP4 | offload | 22.6 GB | – | 32tok/sest. | 1572 tok/s | chat, coding, long documents |
| Qwen3 Coder 30B-A3B Alibaba Qwen | Q4_K_M | offload | 18.9 GB | – | 38tok/sest. | 1203 tok/s | chat |
| Kimi Linear 48B-A3B Moonshot AI | Q4_K_M | offload | 28.9 GB | – | 29tok/sest. | 1272 tok/s | chat, coding, long documents |
| Qwen3 32B Alibaba Qwen | Q4_K_M | offload | 21.3 GB | – | 6tok/sest. | 140 tok/s | – |
| Mistral Small 3.2 24B Mistral AI | Q4_K_M | offload | 15.4 GB | 6k11k with q8 cache | 21tok/sest. | 247 tok/s | chat |
| Gemma 3 27B | Q4_K_M | offload | 17.3 GB | – | 12tok/sest. | 207 tok/s | – |
| DeepSeek R1 Distill Qwen 32B DeepSeek | Q4_K_M | offload | 21.3 GB | – | 6tok/sest. | 146 tok/s | – |
| LFM2.5 24B A2B Liquid AI | Q4_K_M | offload | 15.8 GB | 1k2k with q8 cache | 89tok/sest. | 2311 tok/s | chat, coding |
| Devstral Small 2507 Mistral AI | Q4_K_M | offload | 15.4 GB | 6k11k with q8 cache | 21tok/sest. | 251 tok/s | chat |
| Magistral Small 2506 Mistral AI | Q4_K_M | offload | 15.4 GB | 6k11k with q8 cache | 21tok/sest. | 251 tok/s | chat |
| Magistral Small 2507 Mistral AI | Q4_K_M | offload | 15.4 GB | 6k11k with q8 cache | 21tok/sest. | 251 tok/s | chat |
| Magistral Small 2509 Mistral AI | Q4_K_M | offload | 15.4 GB | 6k11k with q8 cache | 21tok/sest. | 247 tok/s | chat |
| Muse Glimmer 30B meta-models | Q4_K_M | offload | 17.1 GB | – | 12tok/sest. | 190 tok/s | – |
| granite 4.2 30b IBM | Q4_K_M | offload | 19.6 GB | – | 8tok/sest. | 173 tok/s | – |
| Qwen3.8 27B OBLITERATED OBLITERATUS | mlx-4bit | offload | 20.7 GB | – | 7tok/sest. | 184 tok/s | – |
| DeepSeek V4 Flash DeepSeek | UD-Q2_K_XL | won't run | 91.7 GB | – | – | – | – |
| GLM 5.3 Flash Z.ai | UD-Q2_K_XL | won't run | 102.3 GB | – | – | – | – |
| Qwen 3.8 Flash Next Alibaba Qwen | UD-Q2_K_XL | won't run | 74.6 GB | – | – | – | – |
| MiniMax M2.7 MiniMax | UD-Q2_K_XL | won't run | 72.9 GB | – | – | – | – |
| Nemotron 3 Super 120B-A12B NVIDIA | UD-Q2_K_XL | won't run | 52.0 GB | – | – | – | – |
| Qwen3 Coder Next Alibaba Qwen | Q4_K_M | won't run | 46.3 GB | – | – | – | – |
| gpt-oss 120B OpenAI | Q4_K_M | won't run | 59.6 GB | – | – | – | – |
| GLM 4.5 Air Z.ai | Q2_K | won't run | 44.5 GB | – | – | – | – |
| Qwen3 Next 80B-A3B Alibaba Qwen | Q4_K_M | won't run | 46.3 GB | – | – | – | – |
| Llama 3.3 70B Meta | Q4_K_M | won't run | 42.9 GB | – | – | – | – |
| DeepSeek R1 DeepSeek | Q4_K_M | won't run | 378.0 GB | – | – | – | – |
| DeepSeek R1 0528 DeepSeek | Q4_K_M | won't run | 378.5 GB | – | – | – | – |
| DeepSeek R1 Distill Llama 70B DeepSeek | Q4_K_M | won't run | 42.9 GB | – | – | – | – |
| DeepSeek R1 Zero DeepSeek | Q4_K_M | won't run | 378.0 GB | – | – | – | – |
| DeepSeek V3 DeepSeek | Q4_K_M | won't run | 378.0 GB | – | – | – | – |
| DeepSeek V3 0324 DeepSeek | Q4_K_M | won't run | 378.0 GB | – | – | – | – |
| DeepSeek V3.1 DeepSeek | Q4_K_M | won't run | 378.9 GB | – | – | – | – |
| DeepSeek V3.1 Terminus DeepSeek | Q4_K_M | won't run | 378.9 GB | – | – | – | – |
| DeepSeek V3.2 DeepSeek | Q4_K_M | won't run | 378.9 GB | – | – | – | – |
| DeepSeek V4 Flash 0731 DeepSeek | MXFP4 | won't run | 147.2 GB | – | – | – | – |
| DeepSeek V4 Flash Vision Exp DeepSeek | MXFP4 | won't run | 155.9 GB | – | – | – | – |
| Devstral 2 123B Instruct 2512 Mistral AI | Q4_K_M | won't run | 73.4 GB | – | – | – | – |
| GLM 4.5 Z.ai | Q4_K_M | won't run | 205.3 GB | – | – | – | – |
| GLM 4.5V Z.ai | Q4_K_M | won't run | 61.5 GB | – | – | – | – |
| GLM 4.6 Z.ai | Q4_K_M | won't run | 204.5 GB | – | – | – | – |
| GLM 4.6V Z.ai | Q4_K_M | won't run | 67.9 GB | – | – | – | – |
| GLM 4.7 Z.ai | Q4_K_M | won't run | 205.3 GB | – | – | – | – |
| GLM 5 Z.ai | MXFP4 | won't run | 383.9 GB | – | – | – | – |
| GLM 5.1 Z.ai | MXFP4 | won't run | 421.5 GB | – | – | – | – |
| GLM 5.2 Z.ai | UD-Q4_K_M | won't run | 435.4 GB | – | – | – | – |
| GLM 5.3 Z.ai | Q8_0 | won't run | 747.9 GB | – | – | – | – |
| Kimi Dev 72B Moonshot AI | Q6_K | won't run | 63.3 GB | – | – | – | – |
| Kimi K2 Moonshot AI | Q4_K_M | won't run | 579.5 GB | – | – | – | – |
| Kimi K2 Instruct 0905 Moonshot AI | Q4_K_M | won't run | 580.0 GB | – | – | – | – |
| Kimi K2 Thinking Moonshot AI | Q4_K_M | won't run | 580.0 GB | – | – | – | – |
| Kimi K2.5 Moonshot AI | Q4_K_M | won't run | 580.0 GB | – | – | – | – |
| MiniMax M2 MiniMax | Q4_K_M | won't run | 131.6 GB | – | – | – | – |
| MiniMax M2.1 MiniMax | Q4_K_M | won't run | 131.6 GB | – | – | – | – |
| MiniMax M2.5 MiniMax | MXFP4 | won't run | 118.1 GB | – | – | – | – |
| MiniMax M3 MiniMax | MXFP4 | won't run | 241.5 GB | – | – | – | – |
| Mistral Medium 3.5 128B Mistral AI | Q4_K_M | won't run | 73.4 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?
- gpt-oss 20B (MXFP4) fits fully in its memory at 8k context and generates at about 80 tok/s, an estimate from its memory bandwidth. 37 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 92 GB at 8k context, more than the AMD Radeon RX 9070 and system RAM can hold.
- How much does the AMD Radeon RX 9070 cost in Germany?
- We have no current price for the AMD Radeon RX 9070 in Germany.
Speeds on this page are for 8k context. See the best hardware for local AI at every budget.