tonbi(@tonbistudio):In today’s YouTube video, I gave Qwen3.8 27B running locally on my DGX Spark a series of challenging tasks and let it run for days to see how it could handle them. Pretty interesting results! You can see the vision capability pretty clearly I started this gauntlet right after the weights were released, the community has put out more optimized recipes since then, so I think we could squeeze even more out of this little guy! Pushing Qwen3.8-27B to the Limit: How a Small Local LLM Handles Challenging Tasks

tonbi

@tonbistudio

No topic is too difficult, it just hasn't been explained well enough yet Teaching AI on Running

加入 October 2021

1.9K 正在關注 19.4K 粉絲

tonbi@tonbistudio

2026.08.18 17:09

In today’s YouTube video, I gave Qwen3.8 27B running locally on my DGX Spark a series of challenging tasks and let it run for days to see how it could handle them. Pretty interesting results! You can see the vision capability pretty clearly I started this gauntlet right after the weights were released, the community has put out more optimized recipes since then, so I think we could squeeze even more out of this little guy! Pushing Qwen3.8-27B to the Limit: How a Small Local LLM Handles Challenging Tasks