[태그:] Generative AI
-
DeepSeek 4.1 Flash Quietly Matches Frontier AI at a Fraction of Cost
DeepSeek 4.1 Flash reportedly matches Claude Opus-level coding performance while costing far less, thanks to a massive KV cache reduction.
-
GPT-6 Comes to ChatGPT With Interactive “Intelligent UI”
OpenAI rolls out GPT-6 to all ChatGPT users with Intelligent UI, a system that turns answers into interactive diagrams, tools, and apps.
-
AI Agents Port the TypeScript Compiler From Go to Rust
An experimental project called ts-rust used LLM agents to port Microsoft’s TypeScript-Go compiler to Rust, with surprising results on cost and quality.
-
Claude Haiku 5.5: Anthropic’s Fastest, Cheapest Small Model Yet
Anthropic launches Claude Haiku 5.5, a faster and cheaper small AI model built for high-volume tasks, subagent workflows, and agentic coding pipelines.
-
OpenAI Launches Decisions API in Public Beta for Fast Structured Answers
OpenAI’s new Decisions API returns fast, typed answers like probabilities, choices, and scores for classifying and routing content in apps.
-
OpenAI Shares New Math Results From Its Frontier AI Model
OpenAI is publishing new mathematical results from an internal frontier model on GitHub, with Lean proof checks and compute usage data.
-
Dust: A New Way to Train Transformers Without Backpropagation
Researchers introduce Dust, a zeroth-order algorithm that pretrains transformers without backpropagation by perturbing activations instead of weights.
-
Reflection AI’s Beam: A 501B Open-Weight Model for Coding
Reflection AI unveils Beam, a 501B parameter open-weight Mixture-of-Experts model trained with massive-scale reinforcement learning for coding and agentic tasks.
-
ChatGPT Signs Fake New Yorker Cartoons With Real Artists’ Names
ChatGPT is generating New Yorker-style cartoons and signing them with real cartoonists’ pen names, raising new AI attribution and copyright concerns.
-
Kolibri: Aleph Alpha’s Sovereign Open-Weight German AI Model
Aleph Alpha released Kolibri, a 78B open-weight Mixture-of-Experts model with 3B active parameters, built for sovereign, regulated AI deployments.