[MEGATHREAD] Assistant model requests - thread

danielszepi

Hi I’m creating a thread for people to post what providers/models they would want to be implemented/added to the assistant.
I already posted before, but would love to see MiniMax M2.5 (https://www.minimax.io/news/minimax-m25) on Kagi.

Please post the models you would like to see in a format:

Model name:
Provider:
Additional info(links, etc.):

k7absp2h

Thanks for thread

Qwen 3.5 would be great (it's cheap and multimodal)

k7absp2h

Gemini 3.1 Flash Lite Preview and GPT-5.3 Chat

RoxyRoxyRoxy

Honestly good idea, thanks @danielszepi! I'll make the AI team aware of this one 🙂

Rehan

Gemini 3.1 Flash Lite Preview - up
GPT-5.3 Chat - the model listed ChatGPT in our frontend currently points to this
Qwen 3.5 - we only have a single provider that supports this model, and it unfortunately is very slow. When we get other providers that are faster, we'll likely have this up
minimax m2.5 - we're generally careful about the models we offer, as we do not want to clutter the interface with unlimited options. This model seems to be ok in quality, but it isn't clear why one would use it over other models already provided, like Kimi K2.5

kray

Rehan I know minimax m2.5 was already added, but the reason I will use minimax over other model is because of its response style, similar to how people prefer the response of kimi model. Minimax is very pragmatic in its response, so it helps with getting to the information I want faster. Other smarter models tend to explain more information than necessary, causing information clutter. However, minimax m2.5 is very bad at following instructions, so I look forward to minimax m2.7

Rehan

Just a follow up, we'll actually add the minimax model too. It is quite fast, and seems to be popular enough that it is worth adding

k7absp2h

Grok 4.20 Beta

Faster than Gemini 3 Flash
Cheaper than (cost to run) GLM-5

artificialanalysis.ai

cc

Was the o3 pro model removed? Any chance of bringing it back?

k7absp2h

MiMo v2 Flash
Qwen 3.5 Flash
GPT-5.4 Mini
GPT-5.4 Nano

spiffytech

Minimax M2.7 just dropped. I am excited to see M2.5 in the Assistant now though!

mixchiev

There has been new generation of qwen models couple of weeks ago (version 3.5) that shows improvement but kagi assistant show the version 3 still, is there chance to update them?

#qwen #models #updating

k7absp2h

https://kagifeedback.org/d/9948-megathread-assistant-model-requests-thread