Qwen3.6-27B-FP8 vs gemma-3-12b-it
Specs, VRAM requirements and download trends — updated 6 August 2026.
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | Qwen3.6-27B-FP8 | gemma-3-12b-it |
|---|---|---|
| Parameters | 27.8 B | 12.2 B |
| Architecture | Qwen3_5ForConditionalGeneration | Gemma3ForConditionalGeneration |
| Context length | — | — |
| Licence | apache-2.0 | gemma |
| Commercial use | Allowed | Conditional |
| Languages | — | — |
| Downloads 30d | 7,781,120 | 1,303,239 |
| Downloads all time | 23.4 M | 21.6 M |
| Quantizations on the Hub | SAFETENSORS | SAFETENSORS |
| Gated | No | Yes — accept licence |
| First seen on the Hub | 2026-04-21 | 2025-03-01 |
Download trend
Daily snapshots, last 10 days (28 Jul – 6 Aug)
7.8 M1.2 M
Qwen3.6-27B-FP8
gemma-3-12b-it
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
gemma-3-12b-it · fp16
29.1 GB / 24 GB
Qwen3.6-27B-FP8 · fp16
38.6 GB / 24 GB