Skip to content
Search models or paste a model link. Press Enter to search.

Qwen3.6 35B-A3B

Qwen/Qwen3.6-35B-A3B
Left and right arrow keys switch the vendor.
huggingface_int4_embedl-flashhead-mtp_tensorrt-edgellm
Qwen/Qwen3.6-35B-A3B architecture graph
Qwen/Qwen3.6-35B-A3B architecture graph
  • huggingface
  • AGX Thor
Labels
  • Image
  • Text
  • VLM
  • Generation
Build
Embedl-optimized
Precision
W4A16
Runtime
TensorRT Edge-LLM 0.10.1 with Embedl optimizations
Context length
4,096 tokens

Alibaba's mixture-of-experts Qwen3.6: 35B parameters, about 3B active per token, for coding and agent work, with images.

Files

Image Size

Model

docker.io/embedl/model-artifacts:qwen3-6-35b-a3b.huggingface_int4_embedl-flashhead-mtp_tensorrt-edgellm.b951edfb229597a1

21.62 GiB

Runtime

docker.io/embedl/models:runtimes.tensorrt-edgellm.arm64.2026101815014ee0

0.91 GiB

Select hardware to start

Its measured speed and accuracy, its files and how to run it are per device.

Model variants (Pareto)

2 Embedl builds of Qwen3.6 35B-A3B on 2 devices: throughput against Average of 3. Click a build to open it.

84.084.585.085.586.0Average of 3 (%)020406080Throughput (tok/s)No Average of 3 statedBetter ↗