Add model catalog #3941
Add model catalog #3941
Conversation
Docs Preview
|
DouweM
left a comment
DouweM
left a comment
There was a problem hiding this comment.
I think this PR is worth splitting up into 2 PRs because it looks like the page reorg is closer to ready than the popular model catalog
docs/models/alibaba.md
Outdated
| @@ -0,0 +1,55 @@ | |||
| # Alibaba Cloud Model Studio (DashScope) | |||
|
|
|||
| Alibaba Cloud Model Studio (DashScope) provides access to Qwen models via an OpenAI-compatible API. | |||
There was a problem hiding this comment.
If we're going to have this initial paragraph (note that OpenAI, Anthropic, etc do not), we should at least move the link from the configuration section here.
There was a problem hiding this comment.
And let's link "OpenAI-compatible API" to the section that explains those on the openai docs page.
There was a problem hiding this comment.
TODO: review all newly created model pages. Remove initial paragraph (we didn't have it before)
docs/models/azure.md
Outdated
|
|
||
| ## Install | ||
|
|
||
| To use Azure AI Foundry, you need to either install `pydantic-ai`, or install `pydantic-ai-slim` with the `openai` optional group: |
There was a problem hiding this comment.
The "openai" will be confusing/ look likea typo here unless we mention teh "via OpenAI compat API" thing
docs/models/deepseek.md
Outdated
| @@ -0,0 +1,59 @@ | |||
| # DeepSeek | |||
|
|
|||
| DeepSeek provides high-performance AI models with an OpenAI-compatible API. | |||
There was a problem hiding this comment.
Drop the "high-performance" (or the entire paragraph) please, that's too marketing-y for a docs page, and having it here but not elsewhere implies that the other models are not high performance.
docs/models/litellm.md
Outdated
| @@ -0,0 +1,39 @@ | |||
| # LiteLLM | |||
|
|
|||
| LiteLLM provides a unified interface to call 100+ LLM APIs using the OpenAI format. | |||
There was a problem hiding this comment.
consistent wording please, we don't ay "OpenAI format" on the other pages
|
|
||
| <div class="model-card" markdown> | ||
|
|
||
| ### Grok 4 |
There was a problem hiding this comment.
We call this Popular Models and honestly Grok rarely comes up on Slack, so I'd rather stick to the big 3 providers. If we look at usage data, I wouldn't be surprised if Qwen is higher than Grok for example. Maybe we should include Qwen? Then the primary ID can be Alibaba, and we can point out it can be run with Ollama etc
There was a problem hiding this comment.
I mean, I know x-ai gives grok code fast away but yeah people use it a lot. Funnily enough Qwen isn't currently in the ranking. But MiMo (which I never heard of) is.
DouweM
left a comment
DouweM
left a comment
There was a problem hiding this comment.
I think this PR is worth splitting up into 2 PRs because it looks like the page reorg is closer to ready than the popular model catalog
Resolved conflict in docs/models/overview.md: - Kept frontmatter from main (SEO metadata) - Kept our title "Models and Providers" Co-Authored-By: Claude Opus 4.5 <[email protected]>
| ), | ||
| ) | ||
| agent = Agent(model) | ||
| ... |
There was a problem hiding this comment.
should we link here to other models that can also be run in azure to not seem openai-fixated
Addresses review comment - LiteLLM is a proxy server, so pointing directly to OpenAI's API defeats the purpose. Simplified to show the proxy server use case. Co-Authored-By: Claude Opus 4.5 <[email protected]>
- xai.md: grok-2-1212 -> grok-4-fast - vercel.md: claude-4-sonnet -> claude-sonnet-4-5 Co-Authored-By: Claude Opus 4.5 <[email protected]>
result.mp4