Provider changes/OpenAIAI
OpenAI changes and updates.
Evidence-backed changes approved by Auterim’s policy. A provider mention or date alone is not a claim about your own systems.
3 current approved updatesLast verified Nov 4, 2026Approved intelligence
Recent and upcoming changes.
Latest
6FedRAMP endpoint pricing now states a 10% uplift without the previous November 1, 2026 start-date qualifier
FedRAMP endpoint pricing now states a 10% uplift without the previous November 1, 2026 start-date qualifier.Verified Nov 4The o3-deep-research and o4-mini-deep-research models
The o3-deep-research and o4-mini-deep-research models, including their June 26, 2025 snapshots, are deprecated and scheduled to shut down on July 23, 2026.Verified Nov 4The 10% pricing uplift for FedRAMP endpoints now has an explicit effective date of November 1
The 10% pricing uplift for FedRAMP endpoints now has an explicit effective date of November 1, 2026, replacing the previously undated statement.Verified Oct 31Added gpt-6.1-terra pricing for Standard, Batch, and Fast tiers, including cached input and cache-write rates
Added gpt-6.1-terra pricing for Standard, Batch, and Fast tiers, including cached input and cache-write rates.Verified Oct 27The changelog adds the release of GPT-6.1 Terra for the Responses and Chat Completions APIs
The changelog adds the release of GPT-6.1 Terra for the Responses and Chat Completions APIs, describing visual reasoning, computer use, and writing capabilities.Verified Oct 27The Responses API prompt_cache_retention default changes from in_memory to 24h
The Responses API prompt_cache_retention default changes from in_memory to 24h for organizations without Zero Data Retention.Verified Oct 16
Every change page links to OpenAI’s original source. Scan your stack to see likely dependencies, then confirm them in your workspace before Auterim protects them.
Does this affect you?
This page is the same for everyone. Auterim tells you which changes touch your product, and where.