feat: add DeepSeek-R1 model page

[JyotirmoyLaha]

Description

Adds a new model page for DeepSeek-R1, which is currently missing from the guide despite being one of the most significant AI releases of 2025.

Changes proposed

• Added deepseek-r1 entry to pages/models/_meta.en.json
• Created pages/models/deepseek-r1.en.mdx with full model overview including key capabilities, prompting guide, local deployment instructions, and API usage example

Why this matters

DeepSeek-R1 achieved performance comparable to OpenAI o1 on reasoning benchmarks and was one of the most discussed open-source AI models of 2025. The guide currently covers GPT-4, Gemini, Claude, Llama but is missing DeepSeek-R1.

[vercel]

@JyotirmoyLaha is attempting to deploy a commit to the DAIR-AI Team on Vercel.

A member of the Team first needs to authorize it.

[Copilot]

Pull request overview

Adds documentation for the DeepSeek-R1 model to the models guide.

Changes:

• Added a new DeepSeek-R1 model page with overview, prompting guidance, local run instructions, and an API example
• Registered the new page in the English models index (_meta.en.json)

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
pages/models/deepseek-r1.en.mdx	New DeepSeek-R1 documentation page content and examples
pages/models/_meta.en.json	Adds the deepseek-r1 entry so it appears in navigation

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

There is an extra standalone code fence (```), which starts an unclosed code block and causes the trailing content (including the ---) to render/parse as code. Remove the stray fence (and keep `---` only if you really want a horizontal rule).

Suggested change

```

[Copilot]

Screenshot is imported but never used on this page. Please remove the unused import (or add a screenshot using the component) to keep the MDX module clean.

Suggested change

import {Screenshot} from 'components/screenshot'

[Copilot]

The API snippet hard-codes an api_key string. Other model pages use an environment variable (e.g., os.environ.get(...)) to discourage copying secrets into source. Consider switching this example to read the key from an env var and/or mention where to set it.

Suggested change

	client = OpenAI(
	api_key="your_api_key",
	import os
	# Set the DEEPSEEK_API_KEY environment variable with your API key.
	client = OpenAI(
	api_key=os.environ.get("DEEPSEEK_API_KEY"),

[JyotirmoyLaha]

Hi @dair-ai team! 👋

I noticed DeepSeek-R1 was missing from the models section despite being one of the most significant open-source AI releases of 2025. I've added a comprehensive page covering:

• Key capabilities and benchmark performance
• Prompting best practices specific to R1's reasoning behavior
• Local deployment via Ollama
• Official API usage with proper environment variable handling

I've also addressed the Copilot review suggestions. Happy to make any changes based on your feedback!