Update url profiles in assistant.md

[thibaultmol]

pixtral, 3.5 sonnet and llama 3.1 are not working. i don't know what the right formatting is for those.

Feel free to fix it and merge

[david-ros]

@nobodywasishere could you please assist with this

[z64]

• mistral-nemo is Pixtral
• claude-3-sonnet is 3.5 sonnet. We don't deploy both, just the latest.
• llama-3-70b

[nobodywasishere]

@david-ros gpt-4 and gpt-4-turbo should be removed, the rest can stay the same.

[z64]

Might suggest converting this into a table, with the second column having a link to click and try it, copy it, e.g.

Model | ID
-|-
GPT-4o | [gpt-4o](https://kagi.com/assistant?profile=gpt-4o)

So it renders like:

Model	ID
GPT-4o	gpt-4o

[thibaultmol]

@z64 so without the company?
I (personally) would like to keep it there for transparency reasons for the people who don't know.
But maybe add to it:

i added a span to create mouseover text, so that people know the data for those models won't be sent to Meta/Alibaba.

Thoughts?
(should I change company to creator for ex?)

| Company     | Model             | ID                                                   |
|-------------|-------------------|------------------------------------------------------|
| Anthropic   | Claude 3.5 Sonnet | [claude-3.5-sonnet](https://kagi.com/assistant?profile=claude-3-sonnet) |
| Anthropic   | Claude 3 Opus     | [claude-3-opus](https://kagi.com/assistant?profile=claude-3-opus)         |
| Anthropic   | Claude 3 Haiku    | [claude-3-haiku](https://kagi.com/assistant?profile=claude-3-haiku)       |
| OpenAI      | GPT 4o            | [gpt-4o](https://kagi.com/assistant?profile=gpt-4o)                       |
| OpenAI      | GPT 4o Mini       | [gpt-4o-mini](https://kagi.com/assistant?profile=gpt-4o-mini)             |
| Mistral AI  | Mistral Pixtral   | [mistral-nemo](https://kagi.com/assistant?profile=mistral-nemo)           |
| Mistral AI  | Mistral Large     | [mistral-large](https://kagi.com/assistant?profile=mistral-large)         |
| Google      | Gemini 1.5 Pro    | [gemini-pro](https://kagi.com/assistant?profile=gemini-pro)               |
| <span title="Hosted by Firework AI" style="border-bottom: 1px dotted; cursor: help;">Meta</span> | Llama 3.3 70B     | [llama-3-70b](https://kagi.com/assistant?profile=llama-3-70b)             |
| <span title="Hosted by Firework AI" style="border-bottom: 1px dotted; cursor: help;">Meta</span> | Llama 3.1 405B    | [llama-3-405b](https://kagi.com/assistant?profile=llama-3-405b)           |
| <span title="Hosted by Firework AI" style="border-bottom: 1px dotted; cursor: help;">Alibaba</span>        | Qwen QwQ 32b      | [qwen-qwq-32b](https://kagi.com/assistant?profile=qwen-qwq-32b)           |

EDIT: will change it to Fireworks.ai in the final version, just realized the typo

[z64]

@thibaultmol Sure, company is fine too. We usually use the term "Provider" internally, but either works for me. Looks good!

[thibaultmol]

No, you're right Provider makes more sense, didn't think of the word.

Made changes, one more Q: seeing as it's come up various times when people ask what version of 4o for example Assistant is using. Is that something we should be saying in the docs, or is the likely hood of that changing too high that it should be in the UI only (currently not shown, but yeah)

(i'm referring to "gpt-4o-2024-08-06" etc)

[z64]

@thibaultmol

[...] is the likely hood of that changing too high that it should be in the UI only (currently not shown, but yeah)

We do already show it in response metadata;

But maybe you mean adding it someplace like the model picker? Bit tight for space but I can ask our design team what they think. Expect it would be visual noise to most people though, who don't know or care about the versions

But, for the topic at hand, no I would not have it in the docs.

[thibaultmol]

We do already show it in response metadata

sorry yeah, I know. I should have clarified: this is only shown AFTER someone submitted. it should be shown before someone submits

And yeah, something in the picker somehow