AI Systems Status Report - January 2026

This report presents status data for eight major AI systems during January 2026: Anthropic, Cohere, DeepSeek, Google Gemini, Groq Cloud, OpenAI, Perplexity, Replicate, and xAI. The data includes official incidents reported on provider status pages and detection information from IsDown's monitoring system.

Important: Each provider has their own framework for reporting incidents. A provider showing more incidents does not mean they are less reliable - it may indicate more transparent reporting. These providers also vary significantly in size and market share, making direct comparisons inappropriate.

Summary Table

Provider	Official Incidents	Total Incident Time	Peak Incident Day	Typical Resolution
Anthropic	24 (5 MAJOR, 19 MINOR)	~88.8 hours	Wednesday	30-60 min
Cohere	0	0	-	-
DeepSeek	0	0	-	-
Google Gemini	0	0	-	-
Groq Cloud	2 (0 MAJOR, 2 MINOR)	~4.6 hours	Saturday and Tuesday	1-2 hours / 2-4 hours
OpenAI	14 (0 MAJOR, 14 MINOR)	~44.8 hours	Thursday	30-60 min
Perplexity	0	0	-	-
Replicate	0	0	-	-
xAI	0	0	-	-

Provider Status Details

Anthropic

Anthropic reported 24 incidents during January 2026 (5 major, 19 minor), totaling approximately 88.8 hours of incident time with an average resolution time of 221 minutes per incident.

The incidents primarily affected various Claude models including Opus 4.5, Sonnet 4.5, and Haiku 3.5, with elevated error rates being the most common issue type. Several incidents involved authentication and subscription-related problems, particularly affecting Claude.ai and Claude Code services. A notable incident on January 25 involving Opus 4.5 lasted over 30 hours. The platform also experienced issues with structured outputs, API credit purchases, and console performance throughout the month.

Cohere

Cohere reported no incidents during January 2026.

DeepSeek

DeepSeek reported no incidents during January 2026.

Google Gemini

Google Gemini reported no incidents during January 2026.

Groq Cloud

Groq Cloud reported 2 minor incidents during January 2026, totaling approximately 4.6 hours of incident time with an average resolution time of 137 minutes per incident.

The two incidents affected different aspects of the service: a data center failure in Sydney caused model latency issues on January 24, and degraded performance specifically affected the meta-llama/llama-4-scout-17b-16e-instruct model on January 27. Both incidents were resolved within a few hours.

OpenAI

OpenAI reported 14 minor incidents during January 2026, totaling approximately 44.8 hours of incident time with an average resolution time of 192 minutes per incident.

The incidents affected various OpenAI services including ChatGPT, API endpoints, and specific models like GPT-5.1 Codex Max and DALL-E. Common issues included elevated error rates, authentication problems (particularly with Google Authentication on Android), subscription renewal difficulties, and functionality issues with specific features like image processing and GitHub integration. A particularly long-running incident on January 30 affected subscription renewals for over 29 hours.

Perplexity

Perplexity reported no incidents during January 2026.

Replicate

Replicate reported no incidents during January 2026.

xAI

xAI reported no incidents during January 2026.

Conclusion

During January 2026, three of the eight monitored AI systems reported incidents on their official status pages. Anthropic reported the highest number of incidents with 24 total, followed by OpenAI with 14 incidents and Groq Cloud with 2 incidents. The remaining five providers (Cohere, DeepSeek, Google Gemini, Perplexity, Replicate, and xAI) reported no incidents during this period. The most common issues across all providers involved elevated error rates and service degradations, with resolution times typically falling within the 30-60 minute range for most incidents.