BBrainOutput
AMD·GPU

AMD Radeon RX 7900 XT 20GB: Specs & Local-AI Compatibility

20GB RDNA3 card — value memory if ROCm fits your workflow.

Specs

Memory
20 GB
Memory type
GDDR6
Bandwidth
800 GB/s
Approx FP16
103 TFLOPS
Architecture
RDNA 3
Process
TSMC 5nm
Power
315 W
Launch
2022

Models this chip can run

Open models graded for a single AMD Radeon RX 7900 XT 20GB, best fit first.

  • Mistral Small 24B
    Mistral · ~24B · 32K ctx · Apache-2.0

    Fits at Q4_K_M (~14GB) with ~3.6GB headroom — about 1 concurrent instance.

    Q4_K_M · ~14GBRuns well
  • DeepSeek-Coder V2 (class)
    DeepSeek · ~16B · 128K ctx · DeepSeek License

    Fits at Q4_K_M (~11GB) with ~6.6GB headroom — about 1 concurrent instance.

    Q4_K_M · ~11GBRuns well
  • StarCoder2 15B
    StarCoder · ~15B · 16K ctx · BigCode OpenRAIL-M

    Fits at Q8_0 (~17GB) with ~0.6GB headroom — about 1 concurrent instance.

    Q8_0 · ~17GBRuns well
  • Qwen2.5 14B
    Qwen · ~14B · 128K ctx · Apache-2.0

    Fits at Q8_0 (~16GB) with ~1.6GB headroom — about 1 concurrent instance.

    Q8_0 · ~16GBRuns well
  • Qwen3 14B
    Qwen · ~14B · 128K ctx · Apache-2.0

    Fits at Q8_0 (~16GB) with ~1.6GB headroom — about 1 concurrent instance.

    Q8_0 · ~16GBRuns well
  • Phi-3 Medium (14B)
    Phi · ~14B · 128K ctx · MIT

    Fits at Q8_0 (~15GB) with ~2.6GB headroom — about 1 concurrent instance.

    Q8_0 · ~15GBRuns well
  • Phi-4 (14B)
    Phi · ~14B · 16K ctx · MIT

    Fits at Q8_0 (~15GB) with ~2.6GB headroom — about 1 concurrent instance.

    Q8_0 · ~15GBRuns well
  • DeepSeek-R1 Distill 14B
    DeepSeek · ~14B · 128K ctx · MIT

    Fits at Q8_0 (~16GB) with ~1.6GB headroom — about 1 concurrent instance.

    Q8_0 · ~16GBRuns well

Build a private AI Business OS on AMD Radeon RX 7900 XT 20GB

Run your own AI agents on hardware you control — private by design, no per-seat data leaving your premises. BrainOutput helps you pick the right machine and turn it into a working AI Business OS.

Explore the AI Business OS