AI agents: what can they actually do?
From a good answer to a finished task. A practical look at the capabilities, the limits, and the human decisions in between.
Sonnet 5.5 makes everyday work faster
Anthropic emphasizes lower task costs, clearer writing, and practical coding.
Eleven v4 adds more expressive speech
ElevenLabs introduces a general model and a low-latency Turbo variant.
What builders are making with Gemini 3.8 Flash
Google highlights experiments ranging from simulations to interactive engineering models.
Copilot canvases make the workflow a shared interface
GitHub’s walkthrough shows an agent building a board around the task.
Gemini Live gets a visual presence
Google combines real-time dialogue with a generated avatar for enterprise use.
Opus 5.5 combines stronger work with lower pricing
Anthropic reports progress in complex tasks and behavioral evaluations.
GPT-6 Sol and Luna bring more choices to the new family
OpenAI positions two models around different balances of capability and cost.
Isaac ROS 5 brings agents into robotics development
NVIDIA updates its GPU-accelerated packages built on ROS.
Grok Transcribe 2.0 targets difficult real-world audio
The speech-to-text update supports recorded and streaming workflows.
Cohere and Aleph Alpha sign a combination agreement
The proposed company would retain major operations in Canada and Germany.
Grok Build carries project knowledge between sessions
Background notes preserve conventions, decisions, and durable facts.
Perplexity describes putting Astra to work on complete systems
An OpenAI case study looks at testing, software changes, and production monitoring.
Local AI: where to start
Choose the task, the hardware, and the operating boundaries before choosing the model.
North Small Translate focuses on multilingual work
Cohere releases open weights with a non-commercial license.
OpenAI packages cloud-agent orchestration as an API
The managed service brings the Codex harness to longer-running sessions.
GPT-6 Astra opens a new OpenAI generation
The announcement spans coding, computer use, cybersecurity, and science.
Gemini 3.8 Flash extends Google’s agent-focused series
The release pairs a general model with a controlled-access cybersecurity variant.
Gemini learns to inspect video selectively
Google introduces a dynamic alternative to fixed-rate video processing.
Hugging Face brings optimized AI kernels to WebGPU
A browser-focused library includes versioned operations and a hardware test suite.