AI Briefing#llm
The latest AI and developer news about #llm, with the original source and a short summary.
The latest AI and developer news about #llm, with the original source and a short summary.
Mistral released Mistral Small 3.1, an open-source model with multimodal capabilities and a 128k context.
Presents how to build a high-performance, low-power RAG application using FuriosaAI's RNGD accelerator.
Google has released Gemma 3, a new open model supporting multimodality and a 128k context.
OpenAI has released a new set of tools, including the Responses API and Agents SDK, to simplify agent development.
AI21 unveiled Maestro plan orchestration to automate data-intensive enterprise work.
The future of AI reasoning lies not in token prediction but in action planning, the authors argue.
AI21's Maestro plans and executes LLM tasks through requirement verification and iterative refinement.
MCP is a connection protocol that helps AI models access external data and tools in a standardized way.
Kakao CodeBuddy utilizes LLM as a Judge for performance evaluation, proposing methods to enhance evaluation reliability by improving prompt structure and mitigating position bias.
Qwen released **QwQ-32B**, a model that achieves reasoning performance comparable to DeepSeek-R1 through reinforcement learning.
Using Mistral AI's LLM, an agentic workflow was introduced that automatically generates PRDs and development tickets from meeting transcripts.
AI agents are the next-generation AI paradigm that goes beyond the limitations of LLMs to enable autonomous problem-solving.
Japan's secondhand trading platform **Mercari** adopted **GPT-4o mini** to simplify the product listing process and boost sales efficiency.
For enterprise AI, system design and product fit matter more than the LLM alone.
OpenAI has released a research preview of GPT-4.5, featuring broader knowledge and improved emotional intelligence.
OpenAI has released a research preview of GPT-4.5, which scales up unsupervised learning to strengthen knowledge and intuition.
Furiosa has released SDK 2025.1.0, which optimizes LLM inference performance and supports Tool Calling.
LangChain has released OpenEvals, an open-source framework that helps evaluate the performance of LLM and agent applications.
Kanana Flag demonstrated overwhelming performance on Korean benchmarks while reducing training costs by over 50% compared to third-party models.
Qwen has released **QwQ-Max-Preview**, which enhances reasoning and coding performance, and has previewed future app launches and open-sourcing of small models.
AI21 has presented the QLA concept to guarantee accuracy and reliability for enterprise AI.
As the AI industry shifts from training to inference, the importance of efficient inference computing is growing.
Enterprise AI emphasized that trustworthy, controllable, and traceable systems matter more than benchmarks.
Uber is leveraging AI to maximize personalization and operational efficiency across its mobility and delivery services.