Structured Outputs Across Claude, GPT, and Gemini APIs
Compare JSON schema enforcement across Anthropic, OpenAI, and Google: API parameter shapes, SDK tooling, strict constraints, and pricing tiers.
Topic
Models & APIs
How the major model APIs behave in practice, checked against their own documentation.
Claude Sonnet 4.5 Deprecation: Migrating to Sonnet 5.5
Anthropic announced the deprecation of Claude Sonnet 4.5 with retirement set for November 30, 2026. Learn how to update API calls and migrate to Sonnet 5.5.
Gemini responseSchema in Production: What Structured Output Fixed
Lessons from using Gemini JSON mode with responseSchema: minItems for length, required fields, schema descriptions, source_ids validation, Korean keyword counting.
Why LLM Token Counts Differ Between Models (and How to Measure Them)
Same text, different token counts: tokenizer changes like Claude 4.7+'s ~30% more tokens, non-English text, code, and official token counting endpoints.