fix(pricing): align DeepSeek v4 and alias model prices with official docs

[akhilmmenon]

Summary

• Updated pricing/deepseek.json to match official DeepSeek pricing for:
  • deepseek-v4-pro
  • deepseek-chat (deepseek-v4-flash non-thinking alias)
  • deepseek-reasoner (deepseek-v4-flash thinking alias)
• Corrected request/output token rates and cache-read input token rates where values were outdated.

Why

• DeepSeek docs indicate deepseek-chat and deepseek-reasoner are compatibility aliases for deepseek-v4-flash, and current pricing values should reflect the latest official rates.
• This prevents stale cost attribution for DeepSeek models.

Reference

• DeepSeek Models & Pricing

Scope

• Only pricing/deepseek.json was updated.
• No code-path or schema changes.

[Copilot]

Pull request overview

Updates DeepSeek model pricing entries so token and cache-read rates align with the latest official DeepSeek documentation, ensuring downstream cost attribution uses current numbers.

Changes:

• Updated deepseek-chat and deepseek-reasoner request/response token pricing and cache-read input token pricing.
• Updated deepseek-v4-pro request/response token pricing and cache-read input token pricing.

---

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