Crusoe raises $3.9B to build massive data centers and small modular “AI factories”21m ago Google DeepMind launches institute to widen the AGI debate26m ago OpenAI launches Astra for Law, a GPT-6 configuration for legal research30m ago Palantir Hires Former UK Labour Veteran With NHS Deal in Limbo46m ago If METR is overworked, how to alleviate the bottleneck?51m ago Google announces new experimental "CC" AI agent for families52m ago Anthropic Says Claude Drives 26% of Its Research and Development58m ago House heads home to campaign amid calls for urgent AI action59m ago Microsoft exec called AI scraping ‘the largest theft of labor in human history,’ new unredacted filings reveal1h ago PrismML hopes its tiny LLM will change how we all use AI1h ago Meta AI launches Muse personal agent, including apps for iPhone and Mac1h ago Scale AI Reports ROK-FORTRESS Findings on Multilingual AI Safety1h ago Pennsylvania Governor Shapiro Calls for Tougher AI Safeguards1h ago AI can kill us without human extinction: P(Catastrophe)1h ago

Crusoe raises $3.9B to build massive data centers and small modular “AI factories”21m ago Google DeepMind launches institute to widen the AGI debate26m ago OpenAI launches Astra for Law, a GPT-6 configuration for legal research30m ago Palantir Hires Former UK Labour Veteran With NHS Deal in Limbo46m ago If METR is overworked, how to alleviate the bottleneck?51m ago Google announces new experimental "CC" AI agent for families52m ago Anthropic Says Claude Drives 26% of Its Research and Development58m ago House heads home to campaign amid calls for urgent AI action59m ago Microsoft exec called AI scraping ‘the largest theft of labor in human history,’ new unredacted filings reveal1h ago PrismML hopes its tiny LLM will change how we all use AI1h ago Meta AI launches Muse personal agent, including apps for iPhone and Mac1h ago Scale AI Reports ROK-FORTRESS Findings on Multilingual AI Safety1h ago Pennsylvania Governor Shapiro Calls for Tougher AI Safeguards1h ago AI can kill us without human extinction: P(Catastrophe)1h ago

Consumer Tech Business & Enterprise4 min reading time

OpenAI Applies Mitigation as Elevated Error Rates Hit API Models

Unite.AIPublished 2 hours ago

On September 17, 2026, OpenAI experienced elevated error rates affecting 12 API components, including Chat Completions and Images. The company applied a mitigation and monitored recovery, with the incident lasting about an hour.

#api-services #openai #service-reliability

More in Consumer Tech

Consumer Tech2 min read

Google announces new experimental "CC" AI agent for families

Covered by 4 sources3h ago

Consumer Tech2 min read

Meta is reportedly gearing up to launch new smart glasses without a camera

Covered by 4 sources2d ago

Consumer Tech5 min read

Apple Reference Images Explained: The iPhone 18 Pro’s Hardware Solution to AI Slop

Covered by 3 sources2d ago