Qwen 3.8 Flash Next Creative Writing Ability
A user on Reddit's dev_community is experiencing poor creative writing ability with their IQ3_S Qwen 3.8 Flash Next model, despite using a dual MI50 32gb rig. They note that the stories generated are nonsensical, even suggesting Gemma 3 performs better. The user speculates this could be due to Qwen models not excelling in creative writing, a configuration error on their part, or a poor alliterated model choice, as normal Flash Next models are reportedly fine for writing.
时间与来源
时间显示为 UTC
显示时区：UTC
本地时区尚不可用，暂时显示 UTC。
发布当时偏移：UTC+02026年10月29日 07:06 UTC
收录当时偏移：UTC+02026年10月29日 20:00 UTC
- 发布
- 2026年10月29日 07:06
- 收录
- 2026年10月29日 20:00
- 来源类型
- 开发者社区
- 档位
- 社区
- 信源状态
- 同步延迟
档位是按信源手工设定的编辑判断，不是逐条打分。
I use a dual MI50 32gb Rig for an alliterated version of IQ3_S Qwen 3.8 Flash Next. I have seen some people say tha creative writing is good on Flash NeXT but in my experience, it has just been the absolute worst at it.
I’m talking like nothing in the story makes any sense at all. Gemma 3 is better than it imo. Can anyone shed some light on this and see if this is just me not properly configuring it or if it’s just not that great?
Possible reasons that I can think of are just:
- It’s just not good at creative writing as a model, Qwen models have never been that great at it and it’s an agentic model first and foremost.
- It’s a configuration error on my part, some settings I haven’t yet properly set to prioritise creative writing. I will post my settings later, I’m about to go to sleep.
- The alliterated model I chose is just a bad abliteration, normal flash nexts are fine for writing.
The reason I’m making this is because the prose and story in general is just absolutely horrible, just a mess of gibberish, not even a story.