Models › Qwen2.5 Coder 7B
Qwen2.5 Coder 7B: VRAM needs and speed by GPU
We have 2 measured speed results for it.
7.62B parameters · license: apache-2.0 · GGUF files on Hugging Face
Measured speeds
| Hardware | Generate tok/s | Prompt tok/s | Quant | Backend | Test | Source |
|---|---|---|---|---|---|---|
| Apple M2 Ultra | 79.7 | 1,566 | Q8_0 | Metal FA on | tg32 | llama.cpp benches details |
| NVIDIA DGX Spark (GB10) | 29.4 | 2,250 | Q8_0 | CUDA FA on | tg32 | llama.cpp benches details |
Memory needed per quantization
| Quant | File size | Needs (estimate) | Smallest card it fits |
|---|---|---|---|
| Q8_0 | 8.1 GB | 9.1 GB | 12 GB cards |
Expected speed on popular hardware
| Hardware | Best quant that fits | Expected tok/s |
|---|---|---|
| NVIDIA RTX 3060 | Q8_0 | 38.0 estimate |
| NVIDIA RTX 4060 Ti | Does not fit | |
| NVIDIA RTX 3090 | Q8_0 | 81.0 estimate |
| NVIDIA RTX 4090 | Q8_0 | 89.7 estimate |
| NVIDIA RTX 5090 | Q8_0 | 142 estimate |
| AMD Radeon RX 7900 XTX | Q8_0 | 90.1 estimate |
| Apple M4 Pro (20-core GPU) (64 GB) | Q8_0 | 28.6 estimate |
| Apple M3 Max (40-core GPU) (64 GB) | Q8_0 | 41.7 estimate |
| AMD Ryzen AI Max+ 395 (Radeon 8060S) | Q8_0 | 25.7 estimate |
Too big for your card?
Run Qwen2.5 Coder 7B on a rented 24-80 GB GPU by the hour.
Rent a bigger GPU on RunPodAffiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.