Reasoning Models
Reasoning models are a new class of AI that "thinks" before answering: spending extra compute at inference time, building chains of thought and solving problems ordinary LLMs cannot — math, code, science. They kicked off the test-time-compute race. Here is all our coverage of reasoning models: releases, benchmarks (GPQA, ARC-AGI), limits and cost.
Researchers Test 16 Types of Hints for OPSD and Find a Selection Metric
16 privileged-information configs, 3 agentic benchmarks, 3 models: KL metrics predict how much a hint will help OPSD better than directly testing the hint's…
Competence Over Accuracy: Why GPT and Qwen Lose Logic When Problems Get Rephrased
A new methodology called Competence-Based Evaluation (CBE) has uncovered a fundamental flaw in LLM logical reasoning: even strong GPT and Qwen models fail on the…
"Wait" in Reasoning Models Turns Out to Be Inherited from Data, Not Discovered by RL
An analysis of OLMo-3 and MiMo training shows the word "wait" in model reasoning comes from data. RL lifts MATH-500 without it, and marker frequency proves nothing.
Bridged Self-Distillation teaches small models to reason from any teacher, even one behind an API
What the Claude Code leak revealed about the future of Anthropic's AI agents
Meta Unveils Muse Spark: A Multimodal AI Model With Visual Reasoning
OpenVLHarness Gives Multimodal Models Memory and Tools: +9.7–12.6 Points
Hybrid LLMs: Recurrent Layers Suppress Attention, and Staged Training Helps
Researchers describe Competing Learning: in hybrid models, recurrent layers learn retrieval faster and suppress attention. They tested staged training on models up…
Heartbeat Thinking Teaches Voice Agents to Work While You're Still Talking: Latency Down 36.4%
A new scheme for voice agents starts reasoning and tool calls mid-speech. On FDBench v3, tool coverage rose by 16 percentage points and response latency fell by…
Three AI Inflection Points: From ChatGPT to Agents and Physical Intelligence
Moonshot open-sources Kimi K2: a 1-trillion-parameter MoE built for agents
Moonshot releases open Kimi K2 Thinking: 300 tool calls in a row
Moonshot releases Kimi K2.5: a swarm of 100 subagents and 1500 tool calls
Data Rewriting Beats DP-SGD on the Privacy-Quality Trade-Off
A controlled comparison on synthetic medical records with "canaries": fine-tuning leaks the most, DP-SGD hurts quality, and private rewriting offers the best…
AA-Robotics v0.1: GPT-6 Astra controls a robot arm and scores 69 out of 100
Artificial Analysis launched AA-Robotics v0.1, where LLMs control a 6-axis arm. Leader GPT-6 Astra scored 69.4, but it needs 30 minutes and $16.36 per task.