Qwen3 32B
The largest dense model that fits 24GB
Install with Bobbin
Don't have the app? Download it here.
Every parameter is used for every word, which tends to mean more consistent answers on hard questions than a mixture-of-experts model of similar size.
Slower than the 27B, and close to the limit of a 24GB card.
Available versions
| Quantization | Size | VRAM | RAM | Context | |
|---|---|---|---|---|---|
| Q4_K_M | 19.8 GB | 19.0 GBest. | 29.3 GB | 8k | Install |
Memory needed grows with the length of a conversation, so these figures are for the context shown. Anything marked est. is worked out from the file size rather than measured on real hardware — the app says so there too. The desktop app checks all of it against your actual machine before installing.