{
  "id": 10764886,
  "title": "Bring near-Astra intelligence to everyday work with GPT-6.1 Sol on Amazon Bedrock",
  "url": "https://urgent.news/2026/09/29/bring-near-astra-intelligence-to-everyday-work-with-gpt-6-1-sol-on",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-29T19:34:14.000Z",
  "source": {
    "name": "AWS Machine Learning",
    "slug": "aws-machine-learning",
    "url": "https://aws.amazon.com/blogs/machine-learning/bring-near-astra-intelligence-to-everyday-work-with-gpt-6-1-sol-on-amazon-bedrock/"
  },
  "original_language": "en",
  "account": "Amazon has made GPT-6.1 Sol, an advanced AI model, generally available on their Bedrock platform. This upgraded version of GPT-6 Sol brings near-Astra intelligence to everyday professional tasks, particularly in coding, computer use, and software engineering workflows.\n\nGPT-6.1 Sol's enhanced reasoning capabilities allow AI agents to complete tasks more efficiently. For instance, in software engineering, the model can understand unfamiliar repositories, trace dependencies, determine where to make changes, and validate implementations. According to OpenAI, GPT-6.1 Sol matches GPT-6 Astra in DeepSWE v1.1 benchmarking at roughly one-fifth the cost per task, while using less reasoning effort than GPT-6 Sol.\n\nCodex, a tool developed by OpenAI, leverages GPT-6.1 Sol to assist developers throughout the software development cycle. It can work with repositories, local files, terminals, and development tools to write features, fix bugs, and run tests. Developers can access Codex via the desktop app, command line interface (CLI), and supported integrated development environments (IDEs). For AWS users, the Agent Toolkit for AWS connects Codex to AWS documentation, APIs, and services via a single terminal command.\n\nIn addition to coding applications, GPT-6.1 Sol can interpret complex documents, select appropriate tools, and adapt to changing conditions. It can approach GPT-6 Astra in complex document analysis and improve upon GPT-6 Sol's performance in multistep workflows across various business tools. Users can apply these capabilities through ready-to-use experiences or custom-built applications using supported Amazon Bedrock APIs.\n\nHowever, GPT-6.1 Sol also respects limitations and communicates them to users or applications. For instance, it can handle tool failures, restricted actions, or missing information. OpenAI states that GPT-6.1 Sol improves upon GPT-6 Sol in terms of transparency, user intent, and explicit restrictions. Application-level controls allow developers to define available tools and determine responses when an action requires approval or cannot be completed.\n\nTo ensure secure and controlled usage, Amazon Bedrock provides infrastructure and controls for running GPT-6.1 Sol in production. Users can govern model access through AWS Identity and Access Management (IAM) policies, audit invocations using AWS CloudTrail, and use virtual private cloud (VPC) endpoints powered by AWS PrivateLink to keep traffic within network boundaries. Inference runs on hardware-isolated infrastructure with zero-operator access, ensuring that AWS operators cannot access prompts and completions during inference. Data retained by AWS for automated abuse detection is retained for up to 30 days and processed programmatically. Users can request zero data retention through their AWS account team.",
  "summary": "GPT-6.1 Sol is now generally available on Amazon Bedrock, bringing stronger reasoning to coding, computer use, and professional workloads that run frequently.",
  "key_points": [
    "GPT-6.1 Sol available on Amazon Bedrock for everyday work",
    "Enhanced reasoning allows efficient task completion in coding and software engineering",
    "Codex tool leverages GPT-6.1 Sol for developers throughout software development cycle"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}