Amp is extending its AI agent platform beyond software coding tasks by building on several OpenAI offerings. The integration reportedly uses GPT-6 Astra, Ultrafast, GPT-Live-1, and Sign in with ChatGPT to push the fronti...
Google DeepMind announced global availability of Gemini 4 Argon, claiming top results on the Vals Index and an 87% score on the ERQA embodied reasoning benchmark. Google reports a 35% increase in agentic code submissions...
Google has begun a wide rollout of Gemini 4 Argon, its frontier reasoning and agentic coding model, restricted to AI Pro and Ultra subscribers, with Antigravity and AI Studio access limited to Ultra users. The model incl...
Higgsfield has extended its AI video generation from about one minute to more than ten minutes per video. The company credits GPT-6 Astra for longer videos and GPT-Image-2.5 for precise visuals, and says GPT-Live opens a...
Meta's Superintelligence Lab released Muse Wonder and Muse Wonder Full, its first frontier models, which it says match competitors like Fable 5.1 and GPT-6 Astra on benchmarks and lead on the MemoryBench continual-learni...
Google announced a unified Gemini agent that accepts objectives, plans work, uses tools and subagents, and connects to enterprise systems via integrations and MCP. The agent gets its own Workspace account and email, acti...
AI interpretability startup Goodfire has introduced internal activation monitors designed to detect rogue behavior in AI models during inference at significantly reduced operational costs. Deployed via model hosting plat...
Pollo AI has incorporated OpenAI's GPT-5.6, GPT-6 Astra, and GPT-Image-2.5 to assist creators in producing complex imagery and cinematic video commercials. The system is designed to streamline the translation of creative...
Nvidia has expanded its Halos safety platform to robotics, providing dedicated hardware and software guardrails for autonomous machines operating alongside humans. Early adopters including Agility Robotics and Boston Dyn...
French AI startup Mistral has unveiled Mistral Large 4, a 1-trillion-parameter open-weight model nicknamed Le Chonk designed to compete directly with proprietary frontier systems. The model is trained from scratch with s...
