Models › d3
d3: VRAM needs and speed by GPU
At Q4_K_M, d3 needs about 20.9 GB of VRAM. No measured runs yet. Speeds below are estimates. Submit a run.
26.9B parameters · license: apache-2.0 · 262,144-token context · GGUF published Oct 12, 2026 · GGUF files on Hugging Face
Memory needed per quantization
| Quant | File size | Needs (estimate) | Smallest card it fits |
|---|---|---|---|
| Q4_K_M | 19.4 GB | 20.9 GB | 24 GB cards |
| Q8_0 | 28.6 GB | 30.1 GB | 32 GB cards |
| BF16 | 53.8 GB | 55.3 GB | 96 GB cards |
Expected speed on popular hardware
| Hardware | Best quant that fits | Expected tok/s |
|---|---|---|
| NVIDIA RTX 3060 | Does not fit | |
| NVIDIA RTX 4060 Ti | Does not fit | |
| NVIDIA RTX 3090 | Q4_K_M | 33.8 estimate |
| NVIDIA RTX 4090 | Q4_K_M | 37.5 estimate |
| NVIDIA RTX 5090 | Q8_0 | 40.2 estimate |
| AMD Radeon RX 7900 XTX | Q4_K_M | 37.6 estimate |
| Apple M4 Pro (20-core GPU) (64 GB) | Q8_0 | 8.1 estimate |
| Apple M3 Max (40-core GPU) (64 GB) | Q8_0 | 11.8 estimate |
| AMD Ryzen AI Max+ 395 (Radeon 8060S) | Q8_0 | 7.3 estimate |
Too big for your card?
Run d3 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.