PC with RTX 4070/4070 Super or RTX 5070 and 32 GB RAM. Updated 2 November 2026. These machines fit exactly the same models, so they share this page. Their memory speeds range from 504 to 672 GB/s, so words per second scale with memory speed. Speeds shown are for RTX 4070/4070 Super.
PC with RTX 4070/4070 Super or RTX 5070 and 32 GB RAM can run 14 of the 31 models tracked here at a comfortable speed, and another 9 slowly. The largest chat or coding model that fits is Phi-4 14B, which needs about 10.1 GB and generates roughly 25 words a second. The quickest is Gemma 4 E2B at about 113 words a second, which feels faster than you can read.
Video memory (VRAM) is the limit: 12 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 504 GB/s.
Best chat pick: Gemma 4 12B at about 29 words a second. Best coding pick: Qwen2.5 Coder 7B at about 48.
Video models need more memory than this machine has.
Spare memory after loading the model decides how long a conversation or document can be before things slow down. With about 31 GB usable here:
The biggest single jump is Ryzen AI Max+ 395 (Strix Halo), 128 GB, which would unlock 13 more models, including gpt-oss 20B, Gemma 4 26B A4B, Qwen 3.5 35B A3B.
For chat, Gemma 4 12B is the best fit, at about 29 words a second. For coding, Qwen2.5 Coder 7B.
Gemma 4 12B runs at about 29 words a second (runs great); gpt-oss 20B runs at about 19 words a second (slow); Qwen 3.5 35B A3B runs at about 15 words a second (slow).
No. Llama 3.3 70B needs about 45.6 GB and this machine tops out near 31 GB.
| Model | Verdict | Speed | Needs |
|---|---|---|---|
| FLUX.2 klein 4B | Runs great | about 19 s | 10 GB |
| Phi-4 14B | Runs great | 25 words/s | 10.1 GB |
| Gemma 4 12B | Runs great | 29 words/s | 8.9 GB |
| Stable Diffusion XL | Runs great | about 11 s | 8 GB |
| Qwen3 TTS 1.7B | Runs great | about 30 s | 8 GB |
| Gemma 4 E4B | Runs great | 71 words/s | 7.4 GB |
| Qwen 3.5 9B | Runs great | 35 words/s | 7.4 GB |
| Qwen3 VL 8B | Runs great | 38 words/s | 6.9 GB |
| DeepSeek R1 8B | Runs great | 44 words/s | 6 GB |
| Qwen2.5 Coder 7B | Runs great | 48 words/s | 5.4 GB |
| Gemma 4 E2B | Runs great | 113 words/s | 5.3 GB |
| Whisper large-v3 turbo | Runs great | about 42 s | 4 GB |
| Qwen 3.5 4B | Runs great | 69 words/s | 4 GB |
| Kokoro TTS | Runs great | about 6 s | 2 GB |
| Qwen 3.6 35B A3B | Slow | 14 words/s | 24.6 GB |
| Qwen 3.5 35B A3B | Slow | 15 words/s | 23.6 GB |
| Qwen3 Coder 30B | Slow | 16 words/s | 20.4 GB |
| Qwen 3.6 27B | Slow | 3 words/s | 19.9 GB |
| Gemma 4 26B A4B | Slow | 13 words/s | 19.4 GB |
| gpt-oss 20B | Slow | 19 words/s | 15.2 GB |
| ACE-Step 1.5 | Slow | about 3 min | 12 GB |
| Gemma 4 31B | Very slow | 3 words/s | 21.5 GB |
| DeepSeek R1 32B | Very slow | 3 words/s | 21.5 GB |
As an Amazon Associate I earn from qualifying purchases.
Unlocks 9 more: gpt-oss 20B, Gemma 4 26B A4B, Gemma 4 31B, Qwen3 Coder 30B and more.
Unlocks 12 more: gpt-oss 20B, Gemma 4 26B A4B, Qwen 3.5 35B A3B, Gemma 4 31B and more.
Unlocks 13 more: gpt-oss 20B, Gemma 4 26B A4B, Qwen 3.5 35B A3B, Qwen 3.5 122B A10B and more.