Downloadable model · 1 comparison

Qwen3.6-35B-A3B compared

Qwen3.6-35B-A3B has 36 billion parameters, about 18.0 GB of weights at 4-bit and is released under the Apache 2.0 licence. Here it is compared with gemma-4-26B-A4B-it. Pick a matchup for the full side-by-side, a monthly cost calculator and our verdict.

Qwen3.6-35B-A3B at a glance

Accepts: text

Official website

Qwen3.6-35B-A3B next to similar models

The figures that decide whether a downloadable model fits your hardware and your project, for every model it is compared with.

Qwen3.6-35B-A3B compared on size, memory, context and licence
Model	Parameters	Weights at 4-bit	Context	Licence
Qwen3.6-35B-A3B (this model)	36.0B	about 18.0 GB	—	Apache 2.0
gemma-4-26B-A4B-it	25.8B	about 12.9 GB	—	Apache 2.0

Weights only, at 4 bits per parameter. Conversations need memory on top of that; the model hardware calculator works out the full figure for your context length.

gemma-4-26B-A4B-it is the smallest here: its weights need about 12.9 GB at 4-bit, against 18.0 GB for Qwen3.6-35B-A3B.

Compare Qwen3.6-35B-A3B with

vs gemma-4-26B-A4B-it

Qwen3.6 35B-A3B vs Gemma 4 26B A4B

Qwen3.6 35B-A3B and Gemma 4 26B A4B are the mixture-of-experts models in their families: each keeps 26 to 36 billion parameters in memory but uses only 3 to 4 billion for each token, so per-token compute is closer to a small model than a large one. Both are Apache 2.0, both take text and images, and both support 256K-token context. They differ in memory, default thinking behavior and how far they are tuned for agentic coding.

See the comparison

Need a different matchup?

Pick any two API models and we'll build the side-by-side for you.

Build your own comparison