30 noviniek
OpenAI added credit usage analytics — broken down by user, product, and model — plus monthly spend limits configurable per workspace, group, or individual for ChatGPT Enterprise and Edu. The goal: give CFOs and IT clearer visibility into who actually burns the AI budget.
Anthropic shipped new versions of three Claude API tools. code_execution_20260521 surfaces the 90-second per-cell execution limit. web_search_20260318 and web_fetch_20260318 add a response_inclusion parameter to trim consumed result blocks in agentic workflows.
Claude Code 2.1.184–185 adds a `/config key=value` syntax so any option can be set straight from the prompt (thinking mode included), improves stream-stall messaging ('Waiting for API response will retry in'), and extends the silent-retry timer to 20 seconds. Plus subagents, files, auth, and clipboard fixes.
xAI shipped free add-ins for PowerPoint (June 16) and Word (June 18). Generates slides from a prompt, rewrites documents, pulls live data from the web and X. Excel version is in the pipeline.
Vercel ships the security layer for agents — Connect (scoped OIDC tokens), Passport (IdP gating), Sandbox 5 → 24 hours. Plus adapters for Better Auth, Auth.js, eve, AI SDK and MCP.
Claude Code now has /config key=value mid-conversation, macOS sandbox controls, and Apple Events — it can drive Mail, Calendar, Finder. The companion v2.1.183 blocks destructive git commands in Auto Mode.
On June 15, 2026 Anthropic removed `claude-sonnet-4-20250514` and `claude-opus-4-20250514` from the API. Calls now return errors; recommended migrations are Sonnet 4.6 and Opus 4.8.
Alongside its eve framework Vercel shipped infra for long-running agents. Sandbox now runs up to 24 hours, Passport is in public beta, Connect gives agents secure access to external services, and AI Gateway added GLM 5.2.
Google added speech streaming for `gemini-3.1-flash-tts-preview` and announced shutdown dates for legacy media models. Veo retires June 30, Imagen 4 and earlier Gemini image models on August 17, 2026.
Enterprise customers can now plug their own model keys into Copilot CLI. Copilot-authored pull requests now appear in author searches, and auto-generated release notes credit them to the human user.
Linear adds auto-generated changelogs to release pipelines, ships private sub-teams for Business/Enterprise, and rebuilt desktop tabs so each tab keeps its own navigation history.
Demo a task once on a Mac and Codex bundles it into a parameterized skill — no scripting, no RPA, no low-code, just a demo.
Notion launched a Developer Platform — sync any data source into Notion databases via Workers and deploy custom agent tools as hosted endpoints. New: External Agents API (alpha) for Claude/Codex, Notion CLI, Agent SDK, webhook triggers, Markdown API, and MCP improvements that cut token usage by 91% for database operations.
The first xAI model on Bedrock — Grok 4.3 with a 1M context window, configurable reasoning effort, $1.25 in / $2.50 out per MTok. Cheapest US-lab frontier reasoner on Bedrock.
OpenAI launched OpenAI for Healthcare — a product suite for healthcare organizations with HIPAA support — and improved health intelligence in ChatGPT, which now sees 230M+ weekly health-related queries.
Microsoft's small coding model MAI-Code-1-Flash is rolling out across eight Copilot surfaces: CLI, Copilot app, Chat on GitHub, Visual Studio, GitHub Mobile, JetBrains, Eclipse, and Xcode.
The Copilot usage metrics API now exposes an 'ai_credits_used' field with total AI credit consumption per user in both daily and 28-day reports at the enterprise and org levels.
GitHub announced the deprecation of the Opus 4.6 (fast) model in Copilot; users are advised to migrate to current Opus generations or to Microsoft's MAI-Code-1-Flash for fast tasks.
OpenAI made GPT-5.5, GPT-5.4 and Codex available via Amazon Bedrock — ending Azure's hyperscaler exclusivity for OpenAI APIs.
Cursor 3.8 introduces a /automate skill that creates an automation in your local agent session from a plain-language description, plus new GitHub and Slack triggers and computer-use support.
A Claude Code session can now be turned into a sharable, interactive HTML page with live code and multiple data sources — one URL for the team. Repositions Claude Code from assistant to internal-app builder.
The new Claude Design keeps a project aligned with the design system, syncs both ways with Claude Code via /design-sync, and lets you edit directly on the canvas. Imports brand from a GitHub repo, design files, or raw uploads.
The native desktop home for agent-driven development hit GA on all three OSes. Voice conversations via on-device speech-to-text, integrated with session management and github.com.
Fable 5 is the first public-facing version of the Mythos generation; top-tier software engineering, knowledge work and vision, with hard safety guardrails.
Eve is a TypeScript-native open-source framework where every agent = a directory of files (instructions.md + tools/ + skills/). Built in: durable execution, per-agent sandbox, human approvals, evals, OpenTelemetry. Plus the Sandbox runtime now goes up to 24h.
Cursor 3.8 introduces a /automate skill that creates an automation in your local agent session from a plain-language description, plus new GitHub and Slack triggers and computer-use support.
The new Bugbot cuts review time from ~5 minutes to ~90 seconds, finds 10 % more bugs and costs 22 % less per run; adds a /review pre-push flow.
Codex app v26.616 captures a workflow from a single demonstration and generates a natural-language skill description that generalizes to new inputs. Requires Computer Use.
OpenAI made its frontier models and Codex available on AWS — ending Azure's de facto exclusivity for OpenAI API workloads.
Anthropic expanded its Partner Network with a Services Track and a Partner Hub — a channel for integrators and consultancies building on Claude.
