readygpu

AMD Radeon RX 7900 XTX vs AMD Radeon RX 9070 XT

For running models locally, priced in the UK. At 32k context the first runs 54 models fully in memory, the second 33.

In short

The AMD Radeon RX 7900 XTX has more memory for models: 24 GB against 16 GB. At 32k context the AMD Radeon RX 7900 XTX runs 54 of the models we track fully in memory and the AMD Radeon RX 9070 XT 33; only the AMD Radeon RX 7900 XTX holds Qwen 3.8 27B. On the models both run fully in memory, the AMD Radeon RX 7900 XTX is about 1.5× as fast as the AMD Radeon RX 9070 XT (estimated, median across models). In the UK the AMD Radeon RX 7900 XTX costs £780 at eBay (used) and the AMD Radeon RX 9070 XT £751 at eBay: the AMD Radeon RX 9070 XT is £29 cheaper.

Updated . Speeds are estimates; how they are worked out.

AMD Radeon RX 7900 XTX

£780

eBay, used (asking)

Add to build
AMD Radeon RX 9070 XT

£751

eBay

Add to build

Specifications

SpecAMD Radeon RX 7900 XTXAMD Radeon RX 9070 XT
GPURadeon RX 7900 XTXRadeon RX 9070 XT
Memory24 GB GDDR616 GB GDDR6
Memory bus384-bit256-bit
Memory bandwidth960 GB/s640 GB/s
FP16 compute123 TFLOPS195 TFLOPS
SoftwareROCmROCm
Board power355 W304 W
PCIe generationGen 4Gen 5
PCIe lanesx16x16

What each runs

Estimated decode speed with the most precise quant that fits each one est.

ModelAMD Radeon RX 7900 XTXAMD Radeon RX 9070 XT
Qwen 3.8 27B23tok/sQ5_K_Moffload
Qwen 3.6 27B24tok/sQ5_K_Moffload
Gemma 4 31B24tok/sQ4_K_Moffload
Qwen 3.6 35B-A3B122tok/sQ4_K_Moffload
Qwen 3.5 27B24tok/sQ5_K_Moffload
Gemma 4 26B-A4B72tok/sQ6_Koffload
Qwen 3.5 35B-A3B123tok/sQ4_K_Moffload
GLM 4.7 Flash60tok/sQ5_K_Moffload
Devstral Small 2 24B24tok/sQ5_K_Moffload
Nemotron 3.5 Lightning 30B-A3B129tok/sMXFP4offload
gpt-oss 20B98tok/sMXFP465tok/sMXFP4
Qwen3 Coder 30B-A3B63tok/sQ4_K_Moffload
Gemma 4 12B39tok/sQ8_026tok/sQ8_0
Qwen 3.5 9B50tok/sQ8_033tok/sQ8_0
Mistral Small 3.2 24B24tok/sQ5_K_Moffload
Gemma 3 27B24tok/sQ5_K_Moffload
Ministral 3 14B27tok/sQ8_024tok/sQ5_K_M
Qwen3 14B25tok/sQ8_025tok/sQ4_K_M
DeepSeek R1 0528 Qwen3 8B39tok/sQ8_026tok/sQ8_0
Qwen3 8B39tok/sQ8_026tok/sQ8_0
Phi-4 14B24tok/sQ8_0offload
Qwen 3.5 4B94tok/sQ8_063tok/sQ8_0
Ministral 3 8B39tok/sQ8_026tok/sQ8_0
Granite 4.2 8B36tok/sQ8_024tok/sQ8_0
Gemma 4 E4B58tok/sQ8_038tok/sQ8_0
LFM2.5 8B-A1B118tok/sQ8_079tok/sQ8_0
Llama 3.1 8B49tok/sQ6_K32tok/sQ6_K
Qwen2.5 0.5B566tok/sQ8_0377tok/sQ8_0
Qwen3.5 0.8B428tok/sQ8_0285tok/sQ8_0
Qwen3.5 2B284tok/sQ5_K_M190tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B186tok/sQ8_0124tok/sQ8_0
MiniCPM5 2B129tok/sQ8_086tok/sQ8_0
Spark X2.5 4B123tok/smlx-4bit82tok/smlx-4bit
LFM2.5 2.6B108tok/sQ8_072tok/sQ8_0
LFM2.5 VL 3B108tok/sQ8_072tok/sQ8_0
Mellum2.1 12B A2.5B Thinking100tok/sQ8_067tok/sQ8_0
FrogNano 4B 260992tok/sQ8_061tok/sQ8_0
granite vision 4.1 4b84tok/sQ8_056tok/sQ8_0
granite 4.2 3b80tok/sQ8_054tok/sQ8_0
Kimi VL A3B Thinking 250677tok/sQ8_072tok/sQ4_K_M
Kimi VL A3B74tok/smlx-8bit60tok/smlx-6bit
LFM2.5 24B A2B71tok/sQ6_Koffload
GLM 4 9B 041455tok/smlx-5bit37tok/smlx-5bit
DeepSeek R1 Distill Qwen 7B53tok/sQ8_035tok/sQ8_0
GLM 4.6V Flash47tok/sQ8_031tok/sQ8_0
DeepSeek R1 Distill Llama 8B41tok/sQ8_027tok/sQ8_0
GLM 4.1V 9B Thinking40tok/smlx-8bit27tok/smlx-8bit
Ministral 3 8B Reasoning 251239tok/sQ8_026tok/sQ8_0
Devstral Small 250724tok/sQ5_K_Moffload
Magistral Small 250624tok/sQ5_K_Moffload
Magistral Small 250724tok/sQ5_K_Moffload
Magistral Small 250924tok/sQ5_K_Moffload
DeepSeek R1 Distill Qwen 14B24tok/sQ8_0offload
Qwen3.8 27B OBLITERATED23tok/smlx-4bitoffload
Qwen 3.5 122B-A10Boffloadoffload
Qwen3 Coder Nextoffloadwon't run
Mistral Small 4 119Boffloadoffload
GLM 4.5 Airoffloadwon't run
Qwen3 Next 80B-A3Boffloadwon't run
Kimi Linear 48B-A3Boffloadoffload
Qwen3 32Boffloadoffload
DeepSeek R1 Distill Qwen 32Boffloadoffload
granite 4.2 30boffloadoffload
DeepSeek V4 Flashwon't runwon't run
GLM 5.3 Flashwon't runwon't run
Qwen 3.8 Flash Nextwon't runwon't run
MiniMax M2.7won't runwon't run
Nemotron 3 Super 120B-A12Bwon't runwon't run
gpt-oss 120Bwon't runwon't run
Llama 3.3 70Bwon't runwon't run
DeepSeek R1won't runwon't run
DeepSeek R1 0528won't runwon't run
DeepSeek R1 Distill Llama 70Bwon't runwon't run
DeepSeek R1 Zerowon't runwon't run
DeepSeek V3 0324won't runwon't run
DeepSeek V3.1won't runwon't run
DeepSeek V3.1 Terminuswon't runwon't run
DeepSeek V3.2won't runwon't run
DeepSeek V4 Flash 0731won't runwon't run
DeepSeek V4 Flash Vision Expwon't runwon't run
Devstral 2 123B Instruct 2512won't runwon't run
GLM 4.5won't runwon't run
GLM 4.5Vwon't runwon't run
GLM 4.6won't runwon't run
GLM 4.6Vwon't runwon't run
GLM 4.7won't runwon't run
GLM 5won't runwon't run
GLM 5.1won't runwon't run
GLM 5.2won't runwon't run
GLM 5.3won't runwon't run
Kimi Dev 72Bwon't runwon't run
Kimi K2won't runwon't run
Kimi K2 Instruct 0905won't runwon't run
Kimi K2 Thinkingwon't runwon't run
Kimi K2.5won't runwon't run
MiniMax M2won't runwon't run
MiniMax M2.1won't runwon't run
MiniMax M2.5won't runwon't run
MiniMax M3won't runwon't run
Mistral Medium 3.5 128Bwon't runwon't run

Questions

Which is better for local AI, the AMD Radeon RX 7900 XTX or the AMD Radeon RX 9070 XT?
Choose the AMD Radeon RX 7900 XTX to run bigger models. Choose the AMD Radeon RX 7900 XTX for speed. Choose the AMD Radeon RX 9070 XT to spend less.
Which can run bigger models, the AMD Radeon RX 7900 XTX or the AMD Radeon RX 9070 XT?
At 32k context the AMD Radeon RX 7900 XTX runs 54 of the models we track fully in memory and the AMD Radeon RX 9070 XT 33; only the AMD Radeon RX 7900 XTX holds Qwen 3.8 27B.
Which is faster for local AI, the AMD Radeon RX 7900 XTX or the AMD Radeon RX 9070 XT?
On the models both run fully in memory, the AMD Radeon RX 7900 XTX is about 1.5× as fast as the AMD Radeon RX 9070 XT (estimated, median across models).
What is the price difference between the AMD Radeon RX 7900 XTX and the AMD Radeon RX 9070 XT?
In the UK the AMD Radeon RX 7900 XTX costs £780 at eBay (used) and the AMD Radeon RX 9070 XT £751 at eBay: the AMD Radeon RX 9070 XT is £29 cheaper.

More comparisons