What AI can a Radeon RX 9070 XT run?

Quick verdict

PC with Radeon RX 9070 XT (16 GB) and 32 GB RAM can run 16 of the 31 models tracked here at a comfortable speed, and another 9 slowly. The largest chat or coding model that fits is gpt-oss 20B, which needs about 15.2 GB and generates roughly 65 words a second. The quickest is Gemma 4 E2B at about 144 words a second, which feels faster than you can read.

What limits this machine

Video memory (VRAM) is the limit: 16 GB on this card. A model must fit in VRAM to run fast. Anything larger spills into system RAM (32 GB in this setup), which works but is slow, because system RAM moves data at roughly 70 GB/s while the card moves 640 GB/s. AMD cards run most of these tools through ROCm or Vulkan, so setup is a little less smooth than on NVIDIA, and the image and video times here include a 1.5 times penalty.

What it is good for

Chat and coding

Best chat pick: gpt-oss 20B at about 65 words a second. Best coding pick: Qwen2.5 Coder 7B at about 62.

Images

• Stable Diffusion XL: runs great, about about 16 s per image
• FLUX.2 klein 4B: runs great, about about 28 s per image
• Z-Image Turbo 6B: slow, about about 1 min per image

Video

• Wan 2.2 5B: slow, about about 21 min per 4 second clip

Voice and music

• Whisper large-v3 turbo: runs great, about about 42 s per 10 minutes of audio
• Kokoro TTS: runs great, about about 6 s per minute of speech
• Qwen3 TTS 1.7B: runs great, about about 30 s per minute of speech
• ACE-Step 1.5: runs great, about about 1 min per 1 minute song

Fastest models here

• Gemma 4 E2B: about 144 words a second, faster than you can read
• Gemma 4 E4B: about 90 words a second, faster than you can read
• Qwen 3.5 4B: about 87 words a second, faster than you can read
• gpt-oss 20B: about 65 words a second, faster than you can read
• Qwen2.5 Coder 7B: about 62 words a second, faster than you can read

Memory headroom

Spare memory after loading the model decides how long a conversation or document can be before things slow down. With about 34 GB usable here:

• gpt-oss 20B needs 15.2 GB, leaving 19.2 GB spare, plenty for long documents.
• Phi-4 14B needs 10.1 GB, leaving 24.3 GB spare, plenty for long documents.
• Gemma 4 12B needs 8.9 GB, leaving 25.5 GB spare, plenty for long documents.
• Gemma 4 E4B needs 7.4 GB, leaving 27 GB spare, plenty for long documents.
• Qwen 3.5 9B needs 7.4 GB, leaving 27 GB spare, plenty for long documents.

What does not fit, and by how much

• Llama 3.3 70B needs about 45.6 GB and this machine tops out near 34 GB, so it is about 11 GB short.
• gpt-oss 120B needs about 68.8 GB and this machine tops out near 34 GB, so it is about 34 GB short.
• Qwen 3.5 122B A10B needs about 85.5 GB and this machine tops out near 34 GB, so it is about 51 GB short.
• DeepSeek V3 671B needs about 424.7 GB and this machine tops out near 34 GB, so it is about 390 GB short.

How it compares

• PC with RTX 5060 Ti 16GB, RTX 4070 Ti Super, RTX 4080/4080 Super, RTX 5070 Ti, RTX 5080 or RTX 4090 Laptop and 32 GB RAM: the same number of models.
• PC with Radeon RX 7800 XT (16 GB) and 32 GB RAM: the same number of models.
• PC with RTX 3090, RTX 4090 or RTX 5090 Laptop and 32 GB RAM: 7 more models; adds Gemma 4 26B A4B, Gemma 4 31B, Qwen3 Coder 30B and more.
• PC with Radeon RX 7900 XTX (24 GB) and 32 GB RAM: 7 more models; adds Gemma 4 26B A4B, Gemma 4 31B, Qwen3 Coder 30B and more.

Is an upgrade worth it?

The biggest single jump is Ryzen AI Max+ 395 (Strix Halo), 128 GB, which would unlock 11 more models, including Gemma 4 26B A4B, Qwen 3.5 35B A3B, Qwen 3.5 122B A10B.

Questions