docs: Update reasoning_effort parameter mapping for Gemini 2.5 models (v1.17.13)

[roh26it]

Summary

Documentation update for Gateway Enterprise Node v1.17.13:

• Updated reasoning_effort parameter documentation for Gemini/Vertex AI models to reflect the new version-specific behavior:
  • Gemini 2.5 models: reasoning_effort now maps to thinking_budget with specific token allocations
  • Gemini 3.0+ models: reasoning_effort continues to map to thinkingLevel

Token Mapping for Gemini 2.5

reasoning_effort	thinking_budget (tokens)
none	Disabled
low	1,024
medium	8,192
high	24,576

Files Changed

• integrations/llms/vertex-ai.mdx - Updated reasoning_effort section
• integrations/llms/gemini.mdx - Updated reasoning_effort section

Related

• Gateway Enterprise Node v1.17.13 (PR #938)

Test plan

• Verify documentation renders correctly in Mintlify
• Confirm token values match the code implementation