The Monthly

The month

The stories that moved it

The Weekly Reviews' top stories of the month, ranked by how many sources reported them and how much they change. Each links to where it was first said.

1. Claude Sonnet 5.5

   Claude API Release Notes · and 7 more sources

   We've launched Claude Sonnet 5.5 (claude-sonnet-5-5). It's available on the Claude API, Claude in Amazon Bedrock, Claude Platform on AWS, Claude on Google Cloud, and Claude in Microsoft Foundry. For its context window, o…

2. Gemini 4 Argon (gemini-4-argon) generally available (GA)

   Gemini API Changelog · and 12 more sources

   Gemini 4 Argon (gemini-4-argon) generally available (GA): Released Gemini 4 Argon (gemini-4-argon), our most capable model built for complex coding, agentic, math, and science tasks. For API changes, pricing, and migrati…

3. Released GPT-6.1 Sol (gpt-6.1-sol) for complex coding and professional work at a lower cost than GPT-6 Astra.

   OpenAI API Changelog · and 7 more sources

   Feature · gpt-6.1-sol · v1/responses · v1/chat/completions. Released GPT-6.1 Sol (gpt-6.1-sol) for complex coding and professional work at a lower cost than GPT-6 Astra. Standard pricing per 1M tokens for prompts with up…

4. Claude Haiku 5.5 launch

   Claude Release Notes · and 9 more sources

   Claude Haiku 5.5 launchWe just launched Claude Haiku 5.5, the cheapest, fastest, and most capable small model we’ve released, designed for high-volume, cost-sensitive tasks. For more information, see our blog post: Claud…

5. Added Ultrafast mode for GPT-6 Astra in the Responses API.

   OpenAI API Changelog · and 5 more sources

   Feature · gpt-6-astra · v1/responses. Added Ultrafast mode for GPT-6 Astra in the Responses API. Use gpt-6-astra with service_tier: "ultrafast" to reduce the time between generated output tokens. It is available to API c…

6. EmbeddingGemma 2: an open, lightweight multimodal embedding model

   Google DeepMind · and 3 more sources

7. Added Ultrafast mode for GPT-6.1 Sol in the Responses API.

   OpenAI API Changelog · and 3 more sources

   Feature · gpt-6.1-sol · v1/responses. Added Ultrafast mode for GPT-6.1 Sol in the Responses API. Use gpt-6.1-sol with service_tier: "ultrafast" to reduce the time between generated output tokens. It is available to all A…

8. Gemini Nano Banana 2.1 generally available (GA)

   Gemini API Changelog · and 2 more sources

   Gemini Nano Banana 2.1 generally available (GA): Released Gemini Nano Banana 2.1 (gemini-nano-banana-2.1), the latest high-efficiency image generation and conversational editing model. An update to Nano Banana 2 (gemini-…

9. Introducing Mistral Large 4

   Mistral AI · and 2 more sources

10. Added computer use to the Agents API.

    OpenAI API Changelog

    Feature. Added computer use to the Agents API. Agents can complete tasks in an OpenAI-hosted browser, with website access approvals and sign-in handled by your application.

Act on this

What stops working next

Models being shut down and versions reaching end of support in the 90 days after Oct 17, with the replacement each vendor recommends. Only dates they state.

And 37 more deadlines at techgazette.com.br/deadlines/

Security desk

Exploited, and down

What shipped, what rose

Models, and what developers starred

From the papers

Research worth the month

1,882 papers printed this month. A selection, each with its own abstract's words; the week's papers by field are at techgazette.com.br/research.

• BHunter: Generating Blind-Spot Driving Scenarios for Robustness Testing of Perception Systems
  ACM TOSEM · Oct 11
• Distilling Transitional Patterns to Multimodal Large Language Models for Session-based Recommendation
  ACM TOIS (Information Systems) · Oct 11
• Privacy Risks and Leakage Pathways in Agentic and Multi-Agent AI Systems
  ACM TOPS (Privacy and Security) · Oct 10
• ChronoSteer: Bridging Large Language Model and Time Series Foundation Model via Synthetic Cross-Modal Alignment Dataset
  ACM TKDD (Knowledge Discovery from Data) · Oct 11
• Exploring Backdoor Vulnerabilities and Fairness Challenges in Generative Models: A Survey
  ACM Computing Surveys · Oct 12
• One Word Opens the Gate: The Option-Channel Attack on Typed Decision Models as Agent Guardrails
  arXiv cs.AI · Oct 8
• Revealing and Mitigating the Impact of LLM-Generated Content on Retrieval-Augmented Generation Systems
  ACM TKDD (Knowledge Discovery from Data) · Oct 9
• ARC: A Reasoning Recipe for Robot Foundation Models
  arXiv cs.RO · Oct 8
• Reasoning-Informed Visual Editing
  arXiv (vision, retrieval & HCI) cs.CV · Oct 8
• Beyond Spatio-Temporal Priors: A Generalizable Approach for Dense Correspondence Matching
  arXiv cs.CV · Oct 8