OpenAI released the GPT-5.6 family to general availability on July 9 — Sol (flagship), Terra (balanced), and Luna (lowest cost) — after a government-coordinated limited preview begun June 26.
Sol scores 53.6 on Agents’ Last Exam (13.1 pts above Claude Fable 5 at a fraction of estimated cost) and introduces ultra mode, which coordinates multiple agents in parallel for the most complex work.
GPT-5.6 also powers the newly launched ChatGPT Work agent and is now the preferred model in Microsoft 365 Copilot across Word, Excel, PowerPoint, and Cowork.
OpenAI launched GPT-Live, a new full-duplex voice architecture that listens and speaks simultaneously, enabling real-time back-and-forth including active listening cues like ‘mhmm’ without rigid turn-taking.
GPT-Live delegates complex queries to a background frontier model (currently GPT-5.5), keeping the conversation flowing while deeper work completes—decoupling real-time interaction from heavy reasoning.
Two versions—GPT-Live-1 and GPT-Live-1 mini—are rolling out globally to ChatGPT users now; API access is coming soon and developers can sign up for early access.
Anthropic launched Claude Sonnet 5 on June 30 as its most agentic Sonnet yet, closing the performance gap with Opus 4.8 on BrowseComp and OSWorld-Verified while offering better cost efficiency.
It is now the default model for Free and Pro plans, available across Claude Code and the Claude Platform at introductory pricing of $2/$10 per million tokens through August 31, 2026.
Early partners reported it completes multi-step tasks—Salesforce tier updates plus launch emails end-to-end, complex GitHub PRs—where previous Sonnet models would stop short, and proactively checks its own output.
Google DeepMind integrated computer use natively into Gemini 3.5 Flash on June 24, enabling agents that can see, reason, and act across browser, mobile, and desktop environments without a separate model.
Targeted adversarial training mitigates prompt-injection risks; two optional enterprise safeguards allow requiring confirmation for irreversible actions and auto-stopping on indirect injection detection.
Available via the Gemini API and Gemini Enterprise Agent Platform; Browserbase is providing a public demo environment for developers to test immediately.
Meta Superintelligence Labs released Muse Spark 1.1 on July 9 — a multimodal reasoning model with a 1M-token context window, zero-shot MCP generalization, and major gains in tool use, coding, and computer use — priced at $1.25/$4.25 per million input/output tokens.
Alongside it, Meta opened the Meta Model API (public preview, US-only), a closed hosted API that marks a historic strategic pivot away from Meta’s open-weights Llama philosophy toward metered, proprietary inference.
Meta also launched Muse Image (advanced image generation with multi-reference composition and AI effects in Instagram Stories) and previewed Muse Video — all developed by Meta Superintelligence Labs.
OpenAI published data showing Codex now accounts for 99.8% of weekly output tokens generated within the company, with every department—including Legal and Recruiting—switching to agents as their primary AI tool by April 2026.
By May 2026, 80.6% of individual users had made at least one Codex request estimated to require over 30 minutes of human work, and non-developer adoption had grown 137× since August 2025.
The report frames the shift as the beginning of a broader enterprise transformation from single-interaction chatbots to delegated, long-horizon agentic tasks, with coding now routinely done by non-engineers.
Anthropic is moving Claude Fable 5 behind usage-based billing starting July 12 at 11:59 PM PT — subscribers on $20, $100, and $200/month plans will owe additional fees at $10/million input and $50/million output tokens, matching API rates.
This appears to be the first time a frontier AI lab has gated a consumer chatbot model behind per-token billing on top of a subscription, ending the era of unlimited frontier AI access for a flat monthly fee.
A million tokens equals roughly 750,000 words (longer than all of Lord of the Rings), so the fee affects power users and heavy chain-of-thought sessions most; lighter use stays within normal subscription budgets.
US export controls on Claude Fable 5 and Mythos 5 were lifted June 30 after the government accepted Anthropic’s evidence that the reported jailbreak offered no unique capability beyond what Claude Haiku 4.5, GPT-5.5, and Kimi K2.7 could already produce; Fable 5 returned globally July 1. See the original suspension.
Anthropic, Amazon, Microsoft, Google, and Glasswing partners are jointly developing a shared jailbreak severity framework with four criteria: Capability Gain, Breadth of Capability Gain, Ease of Weaponization, and Discoverability — intended to give AI developers a consistent triage standard.
Additional commitments include a new HackerOne program for Fable 5 cyber jailbreak submissions, pre-release government testing access, rapid jailbreak information sharing, and joint development of a common industry evaluation standard.
OpenAI is evolving its GPT-5.5 Bio Bug Bounty into an ongoing private program — the OpenAI Bio Bounty Program — now focused on universal jailbreaks against GPT-5.6 and future frontier models.
The reward for a successful universal jailbreak has been raised from $25,000 to $50,000 for both GPT-5.6 and GPT-5.5; testing for GPT-5.5 ends July 27, after which only GPT-5.6 is in scope.
The program is invite-only via rolling application; participants must sign an NDA, and the expansion signals OpenAI is treating biosafety red-teaming as a permanent, structured part of its safety infrastructure.
OpenAI launched ChatGPT Work on July 9 for Pro, Enterprise, and Edu plans — a Codex-powered agent that handles multi-step tasks across web, desktop, and apps for hours at a time, producing finished sheets, slides, docs, and web apps.
ChatGPT Work integrates a built-in browser, local file access on desktop, Scheduled Tasks for autonomous runs while you’re away, and connections to Microsoft Teams, Slack, and other workspaces.
The Codex desktop app merges into the new ChatGPT desktop app; Work, Chat, and Codex are all available on every plan (including Free) via download on Windows and Mac.
Anthropic launched Claude Tag on June 23 for Slack — Claude joins channels as a persistent team member with shared context, responding when @tagged and proactively surfacing relevant updates in ambient mode.
It is multiplayer (one Claude per channel; anyone can hand off mid-conversation), learns from channel history, and can schedule tasks to run autonomously over hours or days without prompting.
65% of Anthropic’s own product team code is now generated by an internal version of Claude Tag; beta is live for Claude Enterprise and Team customers, with broader platform expansion planned.
OpenAI and Broadcom unveiled Jalapeño, OpenAI’s first Intelligence Processor — a blank-slate LLM inference accelerator designed around OpenAI’s deep understanding of LLM kernels, memory movement, and serving patterns.
Engineered samples are already running GPT-5.3-Codex-Spark at production frequency; early testing shows performance-per-watt substantially better than current SOTA, with a full technical report due in coming months.
The chip is the first in a multi-generation compute platform built with Broadcom and Celestica for gigawatt-scale data center deployment with Microsoft and partners beginning in 2026.
Anthropic published research on July 6 revealing that Claude has a small collection of internal neural patterns — called J-space — that emerged spontaneously during training and function like a conscious ‘global workspace’: Claude can report on them, modulate them on request, and uses them silently for multi-step reasoning.
J-space patterns are causally linked to higher-order cognition (complex reasoning, flexible concept reuse) but are absent from routine tasks like grammar or simple fact recall; blocking J-space leaves fluent interaction intact but removes deliberate reasoning.
The finding, named after the Jacobian-based ‘J-lens’ technique used to read it, mirrors the neuroscientific Global Workspace Theory of consciousness and raises new questions about AI interpretability and machine inner experience.
Anthropic launched Claude Science on June 30 as a unified research environment that integrates PubMed, Jupyter, R, cluster terminals, and HPC login nodes, with 60+ curated skills and connectors for genomics, single-cell analysis, proteomics, structural biology, and cheminformatics.
Every output carries an auditable history — figures include the exact code, environment, and message history that created them; a reviewer agent independently checks citations and calculations.
Claude Science is in beta for Pro/Max/Team/Enterprise users; Anthropic is also offering AI for Science grants (up to 50 projects, up to $30K credits each) with applications open through July 15, 2026.
OpenAI published a systematic audit of SWE-Bench Pro—one of the most widely used coding benchmarks—finding evidence of breaking issues in approximately 30% of its 731 tasks, including overly strict tests, underspecified prompts, low-coverage tests, and misleading prompts that contradict what hidden tests require.
The audit used a data quality pipeline flagging 200 tasks (27.4%), confirmed by human annotation identifying 249 (34.1%); findings were reviewed by five experienced software engineers with disagreements escalated for independent review.
OpenAI advises model developers to treat SWE-Bench Pro results with caution and highlights the growing utility of agents for scalable data quality checks — extending the SWE-bench Verified concerns raised earlier.