---
title: "Comment by Lkzwieder on \"Every Optimized Model for NVIDIA DGX Spark GB10 — Benchmarked & Ranked (April 2026)\" | Hashnode"
description: "AEON-7/Qwen3.6-35B-A3B-heretic-NVFP4 is not working in the real world, is fast but dummy. The regular Qwen3.6-35b-a3b is by far more capable and does real stuff"
url: "https://hashnode.com/posts/optimized-models-nvidia-dgx-spark-gb10-ranked/69fb26b550ecad4533189683/comment/6a1dc932328352c4a3720753"
---

[Discussion](https://hashnode.com/posts/optimized-models-nvidia-dgx-spark-gb10-ranked/69fb26b550ecad4533189683)

Comment [Every Optimized Model for NVIDIA DGX Spark GB10 — Benchmarked & Ranked (April 2026)](https://hashnode.com/posts/optimized-models-nvidia-dgx-spark-gb10-ranked/69fb26b550ecad4533189683)

[Lkzwieder](https://hashnode.com/@lkzwieder)

[Jun 1](https://hashnode.com/posts/optimized-models-nvidia-dgx-spark-gb10-ranked/69fb26b550ecad4533189683/comment/6a1dc932328352c4a3720753)

AEON-7/Qwen3.6-35B-A3B-heretic-NVFP4 is not working in the real world, is fast but dummy. The regular Qwen3.6-35b-a3b is by far more capable and does real stuff. The AEON-7 version needs to mature, there is something that ruins the experience.
