OpenAI launches GPT-5
OpenAI announced the availability of GPT-5, which it says is “smarter across the board” compared to previous models.
Specifically for coding, GPT-5 achieved significant improvement in complex front-end generation and debugging larger repositories. Early testers said that it made better design choices in terms of spacing, typography, and white space, according to the company.
“We think you will love using GPT-5 much more than any previous AI,” CEO Sam Altman said during the livestream. “It is useful. It is smart. It is fast. It’s intuitive.”
Anthropic releases Claude Opus 4.1
This latest update improves the model’s research and data analysis skills, and achieves 74.5% on SWE-bench Verified (compared to 72.5% on Opus 4).
It is available to paid Claude users, in Claude Code, and on Anthropic’s API, Amazon Bedrock, and Google Cloud’s Vertex AI.
The company plans to release larger improvements across its models in the coming weeks as well.
AWS introduces Automated Reasoning checks to reduce AI hallucinations
Automated Reasoning checks are part of Amazon Bedrock Guardrails, and validate the accuracy of AI generated content against domain knowledge. According to AWS, this feature provides 99% verification accuracy.
This was first introduced as a preview at AWS re:Invent, and with this general availability release, several new features are being added, including support for large documents in a single build, simplified policy validation, automated scenario generation, enhanced policy feedback, and customizable validation settings.
Google adds Gemini CLI to GitHub Actions
This new offering is designed to act as an agent for routine coding tasks. At launch, it includes three workflows: intelligent issue triage, pull request reviews, and the ability to mention @gemini-cli in any issue or pull request to delegate tasks.
It is available in beta, and Google is offering free-of-charge quotas for Google AI Studio. It is also supported in Vertex AI and Standard and Enterprise tiers of Gemini Code Assist.
OpenAI announces two open weight reasoning models
OpenAI is joining the open weight model game with the launch of gpt-oss-120b and gpt-oss-20b.
Gpt-oss-120b is optimized for production, high reasoning use cases, and gpt-oss-20b is designed for lower latency or local use cases.
According to the company, these open models are comparable to its closed models in terms of performance and capability, but at a much lower cost. For example, gpt-oss-120b running on an 80 GB GPU achieved similar performance to o4-mini on core reasoning benchmarks, while gpt-oss-20b running on an edge device with 16 GB of memory was comparable to o3-mini on several common benchmarks.
Google DeepMind launches Genie 3
Genie 3 is a frontier model for generating real world environments. It can model physical properties of the world, like water, lighting, and environmental actions.
Users can also use prompts to change the generated world to add new objects and characters or change weather conditions, for example.
According to DeepMind, this research is important because it can enable AI agents to be trained in a variety of simulated environments.
