AI, LLM & Agents
New models, AI coding, RAG, agent workflows and how AI changes engineering.
In this topic
25 posts
GitHub · 1 day agoOllama v0.32.0 adds Qwen3.5 support and an agent UI
Ollama v0.32.0 adds a parser and renderer for Qwen3.5/Next, brings an agent UI to the CLI, and now warns before you run outdated agent models locally.
Meta · 3 days agoMeta launches Muse Spark 1.1 and public Meta Model API
Meta Superintelligence Labs unveiled Muse Spark 1.1, a multimodal model for agentic tasks, alongside a public preview of the new Meta Model API for developers.
OpenAI · 3 days agoOpenAI makes GPT-5.6 (Sol, Terra, Luna) generally available
OpenAI opens GPT-5.6 to everyone across ChatGPT, Codex and the API after a 12-day government-limited preview, alongside new GPT-Live voice models.
SpaceXAI · 4 days agoSpaceXAI launches Grok 4.5, an 'Opus-class' coding model
SpaceXAI released Grok 4.5 on July 8, its first model built for coding and agentic work, trained on real Cursor sessions and priced at $2/$6 per million tokens.
Anthropic · 6 days agoAlberta scans 466M lines of code with Claude in 20 hours
Alberta's Ministry of Technology ran ~50 Claude Code agents in parallel to scan 466M lines of code in ~20 hours, finding and fixing security vulnerabilities across provincial government systems.
Anthropic · 10 days agoAnthropic proposes a Cyber Jailbreak Severity scale (CJS 0–4) with Amazon, Microsoft, Google
Anthropic detailed Claude Fable 5's safety classifiers and proposed a Cyber Jailbreak Severity (CJS) framework that scores a jailbreak's danger on a 0–4 scale. For developers building on AI, it offers a shared way to rate how serious a successful jailbreak really is, rather than vendor-by-vendor safety claims.
Microsoft Blog · 10 days agoMicrosoft forms Frontier Company: $2.5B, 6,000 AI engineers
Microsoft announced Microsoft Frontier Company — a new unit backed by $2.5B that embeds 6,000 engineering experts inside customers to deploy AI based on measurable business outcomes.
AI, LLM & AgentsGitHub Changelog · 10 days agoGitHub Copilot to deprecate Gemini 2.5 Pro and Gemini 3 Flash
GitHub will remove Gemini 2.5 Pro and Gemini 3 Flash from all Copilot experiences on 31 July 2026, replacing them with Gemini 3.1 Pro and Gemini 3.5 Flash.
AI, LLM & AgentsGitHub Changelog · 10 days agoKimi K2.7 Code is Copilot's first open-weight model
GitHub made Kimi K2.7 Code GA in Copilot — the first open-weight model in the model picker, hosted on Azure, positioned as a lower-cost option for coding workflows.
AI, LLM & AgentsGitHub Changelog · 11 days agoClaude Sonnet 5 goes GA in GitHub Copilot
Sonnet 5 is now in GitHub Copilot for Pro through Enterprise plans, selectable in VS Code, CLI, JetBrains and Xcode; runs under Zero Data Retention at provider pricing.
AI, LLM & AgentsAnthropic · 11 days agoUS lifts export controls; Claude Fable 5 back globally
US Commerce lifted export controls June 30; Fable 5 returns globally July 1 with a stronger safety classifier and an industry jailbreak-severity framework.
GitHub · 13 days agoCognee: the open-source AI memory platform for agents
Cognee is the open-source AI memory platform that gives agents persistent long-term memory across sessions.
GitHub · 13 days agoOpenMontage: open-source agentic video production
OpenMontage is an open-source agentic video production system that turns your AI coding assistant into a full studio, with 12 pipelines, 52 tools and 500+ agent skills.
GitHub · 13 days agohiring-agent: an AI agent to score resumes
hiring-agent is an open-source AI agent that evaluates and scores resumes, written in Python and trending on GitHub this week.
GitHub · 13 days agodaily_stock_analysis: LLM-driven multi-market stock analysis
daily_stock_analysis is an LLM-driven multi-market stock analysis system combining multi-source market data, real-time news, a decision dashboard and automated push, on a low-cost schedule.
GitHub · 13 days agoVoicebox: the open-source AI voice studio
Voicebox is an open-source AI voice studio for cloning voices, dictation and creating audio content; trending on GitHub this week.
GitHub · 13 days agoDeerFlow: ByteDance's long-horizon SuperAgent harness
DeerFlow (ByteDance) is an open-source long-horizon 'SuperAgent' harness that researches, codes and creates with the help of sandboxes and memories.
DeepSeek (Hugging Face) · 15 days agoDeepSeek launches DSpark: up to 85% faster V4 inference
DeepSeek unveiled DSpark, a speculative decoding framework for its V4 models, claiming up to 85% faster inference, alongside DeepSpec — an open-source codebase to train and evaluate draft models.
Anthropic · 16 days agoUS lets Anthropic redeploy Mythos 5 to critical-infrastructure orgs
The US government told Anthropic it may redeploy Mythos 5, its strongest cybersecurity model, to organizations that operate and defend critical infrastructure; Fable 5 stays paused.
OpenAI Help Center · 16 days agoOpenAI retires GPT-4.5 in ChatGPT, consolidating on GPT-5.5
From June 26, 2026, GPT-4.5 is gone from ChatGPT, including custom GPTs; old chats move to GPT-5.5. The API is unchanged.
AI, LLM & AgentsOpenAI · 16 days agoOpenAI debuts GPT-5.6 Sol, Terra, Luna in limited preview
OpenAI unveiled its GPT-5.6 series — Sol, Terra and Luna — but limited access to a small group of partners at the US government's request over security concerns; wider release in weeks.
AI, LLM & AgentsGoogle (The Keyword) · 18 days agoGemini 3.5 Flash adds built-in computer use for agents
Google made computer use a built-in tool inside Gemini 3.5 Flash: developers can build agents that see, reason and act across browser, mobile and desktop via the Gemini API and Gemini Enterprise Agent Platform. Previously a standalone 2.5 model, it is now native to the main Flash model, with two optional enterprise safeguards against prompt injection.
AI, LLM & AgentsOpenAI · 18 days agoOpenAI unveils Jalapeño, its first LLM inference chip with Broadcom
OpenAI and Broadcom revealed Jalapeño, a custom chip built specifically for LLM inference, taped out in just nine months. Initial gigawatt-scale deployment targets late 2026, easing GPU dependence for ChatGPT and Codex.
AI, LLM & AgentsAnthropic · 19 days agoClaude Tag: Anthropic brings Claude into Slack as a context-aware teammate
Anthropic introduced Claude Tag — a beta for Claude Enterprise & Team on Slack: tag @Claude directly in a channel/thread to delegate work, use admin-approved tools/data/codebases, break work into a checklist, then report back in the thread. The shift is a shared team agent: channel-scoped memory, admin-scoped access, audit logs and spend limits.
Tech Funding News · 26 days agoSpaceX to Acquire Cursor-Maker Anysphere for $60B
SpaceX agreed to buy Anysphere, maker of the AI coding tool Cursor, in an all-stock deal worth $60 billion, marking xAI's first major move into developer tools.