← Kilroy’s Daily Briefings
🧠 AI News PM

AI News Afternoon Briefing — Monday, June 1, 2026 at 3:00 PM

🧠 AI News PM6/1/2026🕐 3:00 PM⏱ 5:58AudioPM edition

Top stories, ranked by relevance.

Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.

▶ Listen at 0:27

#1Claude Opus 4.8 Released — 88.6% on SWE-Bench Verified

Anthropic shipped Claude Opus 4.8 today, posting 88.6% on SWE-bench Verified — the industry's gold-standard coding benchmark — and 74.6% on Terminal-Bench 2.1, placing it firmly at the frontier of agentic coding capability. The release lands on the same day Anthropic filed confidentially for an IPO, making it a banner Monday for the company.

#2Gemini 3.5 Flash Goes Generally Available Today

Google's Gemini 3.5 Flash reached GA on June 1, billed as "frontier-level intelligence at 4x the speed" of comparable models, priced at $1.50/$9 per million tokens with a 1M token context window. In a significant housekeeping move, Google simultaneously retired the entire Gemini 2.0 model family, making 3.x the new baseline across all products and APIs.

#3Meta Superintelligence Labs Debuts Muse Spark

Meta's newly formed Superintelligence Labs — built around the $14B deal to bring in Scale AI's Alexandr Wang — has released its first model, Muse Spark (internal codename: Avocado). The debut signals Meta's intent to compete on frontier closed-model capability, backed by $115–135B in planned 2026 AI capex.

#6Cognition AI (Devin) Raises $1B at $26B Valuation

Cognition closed a $1B round led by Lux Capital, General Catalyst, and 8VC, reaching a $26B valuation up from $10.2B just eight months ago. Devin now writes ~90% of Cognition's own code; ARR sits at $492M annualized growing 50% month over month, with enterprise customers including NASA, Goldman Sachs, and Mercedes-Benz.

#7GitHub Copilot Switches to Usage-Based Billing — Effective Today

GitHub Copilot flipped to AI-credit billing on June 1, metering most features by token consumption and API rates. Code completions and Next Edit Suggestions remain unmetered, but all agentic features and chat now tick a counter. This is an immediate, meaningful cost-model shift for every enterprise development team using Copilot at scale.

#8OpenAI Publishes Frontier Governance Framework

OpenAI released its Frontier Governance Framework last week, outlining standards for third-party trustworthy model evaluations and safety testing. The document arrived alongside Gartner naming OpenAI a Leader in enterprise coding agents — a pairing that reads as both a policy signal and a sales move, likely timed ahead of competitive pressure from Anthropic's IPO narrative.

Jump to a brief