OpenAI API
- OpenAPI contract
Vendor coverage
OpenAI API changes & breaking changes
OpenAI's API surface moves as fast as its models: new endpoints, new parameters, response schemas that grow and shift between releases. Detektu diffs the published OpenAPI specification so your integration keeps up.
Last checked Oct 20, 2026, 3:02 PMLast change detected Oct 20, 2026, 9:01 AM
Detected changes
Latest OpenAI API changes
Every entry is a change Detektu detected in OpenAI's published contracts, with exact before/after evidence on its page.
- InfoRequest body any of added: request-body-any-of-added
openapi:POST /organization/projects/{}/api_keys/{}::request.request-body-any-of-added - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/web_search_calls::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/vector_stores::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/moderations::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/images::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/file_search_calls::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/embeddings::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/decisions::response.200.application.json.data.items.results.items.input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/completions::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens - InfoResponse optional property added: input_cache_write_30m_tokens
openapi:GET /organization/usage/code_interpreter_sessions::response.200.application.json.data.items.results.items.oneOf[#.components.schemas.UsageCompletionsResult].input_cache_write_30m_tokens
Coverage
What Detektu monitors
Official products and contracts checked for changes.
Change model
How OpenAI ships changes
The release patterns that shape your integration risk.
OpenAI publishes an OpenAPI specification covering its API surface: chat completions, responses, embeddings, files, and more. It changes often, mostly additively, with occasional sharp edges. Parameters change shape, and response fields appear or move as capabilities roll out.
Detektu diffs every revision of the spec and classifies each change by severity, so a new optional parameter reads as signal while a changed response schema on an operation you call escalates with exact before/after evidence.
Monitored for you
Do OpenAI's changes affect your integration?
Choose the operations and fields your services use. Detektu matches OpenAI findings to that declared usage so your team can review the relevant evidence. No agent or repository access required.