Reading Time: 5 minutesOpenAI has previewed GPT-5.6 as a three-model family — Sol for maximum power, Terra for everyday balance, and Luna for speed and lower cost. The rollout is phased, with safeguards built around Sol’s enhanced cybersecurity capabilities and a Cerebras partnership targeting up to 750 tokens per second by July.
Reading Time: 3 minutesClaude Code version 2.1.197 makes Claude Sonnet 5 the new default model with a 1-million-token context window and promotional pricing through August 31. The release also brings organisation-level model controls, significant background-agent reliability improvements, a 37% streaming CPU reduction, and a critical MCP security fix that prevents untrusted repositories from self-approving their own tools.
Reading Time: 5 minutesClaude Sonnet 5 launched on 30 June 2026 with near-Opus 4.8 performance, a 1 million token context window, and adaptive thinking enabled by default — but a new tokenizer makes English-language inputs roughly 30 to 42 percent more expensive in real terms than the headline price suggests. Simon Willison’s analysis documents the tokenizer gap empirically, and Indian professionals using Claude-powered tools should audit their costs before the introductory discount expires on 31 August 2026.
Reading Time: 5 minutesAnthropic has launched Claude Science, a flagship agentic AI product designed for scientific research and drug discovery, positioning itself to challenge Google DeepMind’s decade-long dominance in AI for science. The product is now available to paid Claude subscribers and Anthropic plans to use it for its own drug discovery research into neglected diseases.
Reading Time: 4 minutesAI agents drift over time as models update and business data evolves, making ongoing recalibration essential rather than optional. Zapier’s guidance outlines how to sharpen instructions, build review loops, and version your prompts to maintain reliable agent performance.
Reading Time: 5 minutesOpenAI and Broadcom have unveiled Jalapeño, a custom ASIC designed in roughly nine months that will let OpenAI reduce its dependence on Nvidia GPUs and move toward full-stack infrastructure control. The chip is slated for deployment later in 2026 and full scaling through early 2028, timed strategically with OpenAI’s IPO ambitions.
Reading Time: 5 minutesOpenAI launched GPT-5.6 Sol — its strongest model yet for coding, biology, and cybersecurity — but access is gated behind a U.S. government-requested trusted-partner review. The Neuron reports that the release process, not the benchmarks, is the defining story of this launch.
Reading Time: 5 minutesProsecutors in the Palisades wildfire arson trial used the defendant’s ChatGPT conversation logs — including fire image requests, anger rants, and queries about fire liability — as criminal evidence. The case is a landmark moment revealing that AI chatbot histories are legally accessible and may be used to establish intent in criminal proceedings.
Reading Time: 5 minutesFernando Irarrázaval’s public challenge saw 2,000 people send 6,000 emails trying to extract secrets from an AI assistant, and none succeeded — a result that Simon Willison’s analysis calls consistent with broader improvements in frontier model training against prompt injection. However, Willison cautions that failed casual attempts offer no guarantee against a sophisticated, targeted attacker, and the post explains what this means for Indian professionals deploying AI in real business workflows.
Reading Time: 5 minutesOpenAI has unveiled GPT-5.6 — a three-model suite comprising Sol, Terra, and Luna — less than 24 hours after reports of Trump administration pressure to stagger its release. The flagship Sol model is priced at $5 input / $30 output per million tokens and targets coding, cybersecurity, biology, and long-horizon agentic AI tasks.