Time with problems

• Cursor1 h 37 min
• Claude1 h 27 min
• Claude Code1 h 27 min
• Claude API1 h 27 min
• OpenAI API1 h 22 min
• ChatGPT1 h 18 min
• Fireworks AI1 h 1 min

Total time a service had confirmed problems during the day (UTC).

Incidents

• Anysphere: Degraded agent personalization and codebase sync

  3 h 17 min

• Anysphere: Payment System Sync Issues

  5 min

• Fireworks AI: Service Degradation for Kimi K3

  30 min

• Anthropic: Elevated errors for Claude Sonnet 5.5

  42 min

• Fireworks AI: Service Degradation for GLM 5.3 Flash

  24 min

• OpenAI: 6.1 Sol Degraded Model Performance

  1 h 18 min

• Anthropic: Elevated errors for Claude Sonnet 5.5

  41 min

• Anysphere: Delayed responses in Grok Bot

  10 min

New models (2)

• OpenAI GPT-6.1 Terra Pro$1.00 / $5.00
• PrismML Ternary Bonsai 2 9B$0.08 / $0.13

Price changes

• Tencent Hy4 preview$2.25 → $2.50 +11%
• Tencent Hy3$0.33 → $0.528 +60%
• DeepSeek V4.1 Flash$0.01 → $0.42 +4100%
• Qwen3.8 27B$2.50 → $1.25 -50%
• DeepSeek V4 Flash 0423$0.182 → $0.28 +54%
• Xiaomi MiMo-V2.5$0.238 → $2.00 +740%
• Z.ai GLM 5.1$4.40 → $3.04 -31%
• Qwen3.5-27B$2.16 → $2.60 +20%
• Qwen3.6 35B A3B$0.90 → $0.95 +6%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• Xiaomi MiMo-V2.5-Pro$0.609 → $1.80 +196%
• DeepSeek V4.1 Flash 1017$0.60 → $1.20 +100%
• DeepSeek V4 Pro 0813$1.98 → $3.96 +100%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• Z.ai GLM 5.3$1.40 → $1.40 0%
• MoonshotAI Kimi K3$20.00 → $20.00 0%
• Qwen3.8 27B$1.25 → $2.50 +100%
• DeepSeek V4 Flash 0731$1.32 → $0.238 -82%
• DeepSeek V4.1 Flash 1017$1.20 → $0.60 -50%
• Z.ai GLM 5.3 Flash$0.50 → $0.076 -85%
• and 107 more

Retirements announced

• Z.ai GLM 4.5Vretires 2026-12-31
• Meta Llama 3.1 70B Instructretires 2026-11-03
• Z.ai GLM 4.5Vretires 2026-12-31

Worth reading

• ★Report: Anthropic trained Claude on illegally obtained Stanford booksHacker News · 540
• ★Amazon Is Gutting Its AI Division After Sustained FailureHacker News · 332
• ★OpenAI expands Oracle partnership to integrate apps into ChatGPTOpenAI
• Wikipedia's Cleanup of AI-Generated ContentHacker News · 484
• Police Use AI to Stop Drivers for Suspicious BehaviorHacker News · 378
• Early progress on accelerating drug development with ClaudeHacker News · 264
• Check Point uses OpenAI Daybreak to accelerate vulnerability researchOpenAI
• Together AI introduces auto-repair for GPU clustersTogether AI
• Retell uses OpenAI models for 100M monthly callsOpenAI
• Ollama v0.40.12 adds LFM3 and d1 models to MLX backendOllama releases
• LFM3 for specialized multimodal modelsHugging Face
• Discussion on AI increasing workload despite promises of efficiencyHacker News · 649
• Discussion on The Mythical Man-Month in the Age of AIHacker News · 462
• Survival LLM: Local AI Assistant and Emergency LibraryHacker News · 416
• Ollama releases v0.40.12 with Gemma 4 and llama.cpp updatesOllama releases
• Can we trust LLMs to judge whether a research idea is novel?Hugging Face