Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
1–10 of 482 posts
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#2A bit skeptical about a 27B model comparable to opus...
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#3A bit skeptical about a 27B model comparable to opus...
you'd be surprised how good small models have gotten. Size of the model isnt all that matters.
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#4A bit skeptical about a 27B model comparable to opus...
you'd be surprised how good small models have gotten. Size of the model isnt all that matters.
My experience with qwen-3.6:35B-A3B reinforces this, gonna give this a spin when unsloth has quants available
Gemini flash was just as good as pro for most tasks with good prompts, tools, and context. Gemma 4 was nearly as good as flash and Qwen 3.6 appears to be even better.
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#5Has anyone tested it at home yet and wants to share early impressions?
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#6Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#7A bit skeptical about a 27B model comparable to opus...
Some of these benchmarks are supposedly easy to game. Which ones should we pay attention to?
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#8A bit skeptical about a 27B model comparable to opus...
For at least a year now, it has been clear that data quality and fine-tuning are the main sources of improvement for mediym-level models. Size != quality for specialized, narrow use cases such as coding.
It’s not a surprise that models are leapfrogging each other when the engineers are able to incorporate better code examples and reasoning traces, which in turn bring higher quality outputs.
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#9Has anyone tested it at home yet and wants to share early impressions?
I have been kicking the tires for about 40 minutes since it downloaded and it seems excellent at general tasks, image comprehension and coding/tool-calling (using VLLM to serve it). I think it squeaks past Gemma4 but it's hard to tell yet.
Re: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
#10Earlier quoted context omitted.
you'd be surprised how good small models have gotten. Size of the model isnt all that matters.
My experience with qwen-3.6:35B-A3B reinforces this, gonna give this a spin when unsloth has quants available Gemini flash was just as good as pro for most tasks with good prompts, tools, and context. Gemma 4 was nearly as good as flash and Qwen 3.6 appears to be even better.
> when unsloth has quants available