AI Briefing#agent
The latest AI and developer news about #agent, with the original source and a short summary.
The latest AI and developer news about #agent, with the original source and a short summary.
Kanzhun released the 33B dense model Nanbeige4.3-33B under the Apache 2.0 license.
The 33B model serves as a teacher for upcoming smaller agents and outperforms Qwen3.8-27B on Terminal-Bench 2.1 but trails on SWE-Bench Pro.
NVIDIA has released Nemotron 3.5 Bolt E4B, a small model for interactive agents that operates on 4GB VRAM.
Achieved an AA Index score of 24.1 with an MoE architecture that activates only 3B of its total 35B parameters
The Qwen3-based CLM supports all of Jev's features and achieved 87.6% performance on Terminal-Bench.
Xiaomi released 'MiMo-V2.6-Distill-Qwen-9B', an agent-specialized model fine-tuned from Qwen3.5-9B using MiMo data.
The open-source MCP app 'Chat On Steroids' has been released, enabling local file modifications and terminal command execution via ChatGPT conversations.
Human-Agent-Society has released 'Reef', an open-source infrastructure that updates model weights and harnesses in real time using user feedback during agent serving.
Open-sourced a web-search Research agent and a document chatbot Support agent.
Cloudflare unveiled Artifacts, a version-controlled filesystem built for the agent era.
openclaw 4.14 conflicts with lossless-claw, causing the agent to freeze.
A new Agent tab has appeared in Gemini Enterprise, revealing a task-oriented workspace.
Google appears set to soon expand reusable Skills features across Gemini and AI Studio.
LLMs can only perform real work when there is a harness and layered design in place.
Yogiyo built a DDL review bot with Amazon Bedrock, improving review speed by about 20%.
Continue rolled out 5 features—including CLI, Next Edit, and Plan Mode—over 5 consecutive days, strengthening its developer-centric AI workflow.
They implemented an in-house Agent that finds Confluence documents via natural language, using workflows and CQL.
The upgraded Claude 3.5 Sonnet scored 49% on SWE-bench Verified.
Alongside the v0.10 release, it introduces new integrations and learning resources for RAG and agents.
DeepSeek-V3.1-Terminus improves language consistency and agent performance.