The AI Bench
24 GB · 960 GB/s

What can the AMD Radeon RX 7900 XTX run?

24 GB with RTX 4090-class bandwidth. Best if you already own one.

Best models for each task

Our top three for this machine. Chat and image tasks assume an 8K-token conversation; coding, documents and agents assume 32K.

Chat & writing18 of 25 run

  1. Qwen3.8 27BRuns wellQ4_K_M · needs 19.9 GB at 8K · 19–30tok/sSmooth
  2. Gemma 4 31BRuns wellQ4_K_M · needs 21.4 GB at 8K · 18–30tok/sSmooth
  3. Muse Glimmer 30BRuns wellQ4_K_M · needs 19.8 GB at 8K · 20–30tok/sSmooth

Coding20 of 26 run

  1. Qwen3.8 27BRuns wellQ4_K_M · needs 21.5 GB at 32K · 18–30tok/sSmooth
  2. Muse Glimmer 30BRuns wellQ4_K_M · needs 20.2 GB at 32K · 19–30tok/sSmooth
  3. Qwen3.6 35B-A3BRuns wellUD-IQ4_XS · needs 20.3 GB at 32K · 75–150tok/sFast

Long documents13 of 20 run

  1. Qwen3.8 27BRuns wellQ4_K_M · needs 21.5 GB at 32K · 18–30tok/sSmooth
  2. Qwen3.6 35B-A3BRuns wellUD-IQ4_XS · needs 20.3 GB at 32K · 75–150tok/sFast
  3. Gemma 4 12BRuns greatQ8_0 · needs 14.7 GB at 32K · 25–40tok/sSmooth

Agents & tools20 of 26 run

  1. Qwen3.8 27BRuns wellQ4_K_M · needs 21.5 GB at 32K · 18–30tok/sSmooth
  2. Qwen3.6 35B-A3BRuns wellUD-IQ4_XS · needs 20.3 GB at 32K · 75–150tok/sFast
  3. Muse Glimmer 30BRuns wellQ4_K_M · needs 20.2 GB at 32K · 19–30tok/sSmooth

Images in9 of 13 run

  1. Qwen3.8 27BRuns wellQ4_K_M · needs 19.9 GB at 8K · 19–30tok/sSmooth
  2. Qwen3.6 35B-A3BRuns wellUD-IQ4_XS · needs 19.8 GB at 8K · 90–170tok/sFast
  3. Muse Glimmer 30BRuns wellQ4_K_M · needs 19.8 GB at 8K · 20–30tok/sSmooth

Make images6 of 6 run

  1. Stable Diffusion XL Base 1.0Runs greatFP16 · needs 7.9 GB
  2. FLUX.2 [klein] 4BRuns greatBF16 · needs 14.0 GB
  3. Z-Image-TurboRuns greatBF16 · needs 17.0 GB

Speak text7 of 7 run

  1. Qwen3-TTS 1.7BRuns greatBF16 · needs 5.5 GB
  2. Kokoro-82MRuns greatFP32 · needs 1.4 GB
  3. Chatterbox Multilingual V3Runs greatFull · needs 4.2 GB

Transcribe7 of 7 run

  1. Voxtral Mini 4B RealtimeRuns greatBF16 · needs 17.0 GB
  2. Whisper large-v3-turbo (via faster-whisper)Runs greatFP16 · needs 2.6 GB
  3. Nemotron 3.5 ASR Streaming 0.6BRuns greatFP32 · needs 3.5 GB

Biggest model it runs well

Qwen3.6 35B-A3B (UD-IQ4_XS, 17.7 GB file) at about 90–170tok/sFast.

Should you buy it?

Buy it if

You run Linux and want 24 GB from AMD.

Skip it if

New prices are close to used NVIDIA 24 GB cards with easier software.

Watch out for

  • End of life: mostly old stock and used cards now.
  • Read our AMD guide first.