Can I Run OpenAI gpt-oss-20b on an RX 7800 XT?

Yes. At Q4_K_M and an 8K context it needs about 14.0 GB, which is 88 percent of the RX 7800 XT’s 16 GB.

Check your own setup with the Can My GPU Run This AI Model? calculator. The numbers below come from the same data.

Memory needed on the RX 7800 XT

Quantization Weights Total at 8K (est.) On 16 GB Speed (est., tok/s)
Q4_K_M 12.8 GB 14.0 GB Fits –
Q5_K_M 14.9 GB 16.1 GB No –
Q6_K 17.3 GB 18.5 GB No –
Q8_0 22.2 GB 23.4 GB No –

The highest quantization that still fits comfortably is Q4_K_M.

Context length matters

Context KV cache Total at Q4_K_M (est.) On 16 GB
4K 0.10 GB 13.9 GB Fits
8K 0.20 GB 14.0 GB Fits
16K 0.41 GB 14.2 GB Fits
32K 0.81 GB 14.6 GB Fits

What speed to expect

OpenAI gpt-oss-20b is a Mixture-of-Experts model, so we do not estimate speed from bandwidth.

Estimates: weights = parameters × bits per weight; KV cache from the model’s published architecture; about 1 GB of runtime overhead. “Fits” means at most 92 percent of VRAM, “Tight” is 92 to 100 percent. Speed is memory bandwidth divided by model size at 60 to 80 percent efficiency. These are planning estimates, not our own benchmarks.

Verdict

Run it at Q4_K_M and enjoy. For a wider view, see what GPU you need for OpenAI gpt-oss-20b and everything the RX 7800 XT can run.

Frequently asked questions

Can the RX 7800 XT run OpenAI gpt-oss-20b at Q8?

Q8_0 needs about 23.4 GB at 8K context: no on 16 GB.

How much context can the RX 7800 XT handle with OpenAI gpt-oss-20b?

At Q4_K_M, 32K context needs about 14.6 GB in total (fits on 16 GB).

Sources and method

Last updated 10 Nov 2026.

Spynium


Leave a Comment