AllLLMRoboticsStartupsResearchEthicsBusinessHardwareAutonomousMedicineCreative
Topic hub

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.

Official site ↗290 articles
In the glossary:Reasoning ModelChain-of-Thought (CoT)Test-Time ComputeTree of Thought
Reasoning Drift in AI: How a Model Degrades While the Service Keeps Running
LLM·Habr AI

Reasoning Drift in AI: How a Model Degrades While the Service Keeps Running

Oct 18, 2026·5 min
Three AI Inflection Points: From ChatGPT to Agents and Physical Intelligence
LLM

Three AI Inflection Points: From ChatGPT to Agents and Physical Intelligence

Oct 18, 2026·3 min
Moonshot open-sources Kimi K2: a 1-trillion-parameter MoE built for agents
LLM

Moonshot open-sources Kimi K2: a 1-trillion-parameter MoE built for agents

Oct 17, 2026·3 min
Moonshot releases open Kimi K2 Thinking: 300 tool calls in a row
LLM

Moonshot releases open Kimi K2 Thinking: 300 tool calls in a row

Oct 17, 2026·2 min
Moonshot releases Kimi K2.5: a swarm of 100 subagents and 1500 tool calls
LLM·Moonshot Kimi Blog

Moonshot releases Kimi K2.5: a swarm of 100 subagents and 1500 tool calls

Moonshot unveils the open multimodal Kimi K2.5: further trained on 15T tokens, with a swarm of up to 100 subagents and up to 4.5x faster execution on complex tasks.

Oct 17, 2026·3 min
Moonshot open-sources Kimi K2.6: 300 sub-agents and 13 hours of autonomous coding
LLM·Moonshot Kimi Blog

Moonshot open-sources Kimi K2.6: 300 sub-agents and 13 hours of autonomous coding

Moonshot open-sourced Kimi K2.6: the model spent 13 hours tuning an exchange engine, sped up Zig inference from 15 to 193 tokens/s, and scales its swarm to 300…

Oct 17, 2026·3 min
Xiaomi-Lens-0, a 4B model, beats Qwen3.5 by 5.6 points on RefCOCO
LLM·Hugging Face Daily Papers

Xiaomi-Lens-0, a 4B model, beats Qwen3.5 by 5.6 points on RefCOCO

Xiaomi releases Lens-0: a 4B model built on Qwen3.5 that draws masks from its own visual latents, gains 5.6 points on RefCOCO and leads on cIoU.

Oct 16, 2026·2 min
Stanford Posts LLM Training Lecture: 16 Chapters From Pretraining to GRPO
LLM·YouTube: Stanford Online

Stanford Posts LLM Training Lecture: 16 Chapters From Pretraining to GRPO

Stanford's third CME295 lecture walks through the full LLM training stack in 16 chapters: pretraining, SFT, LoRA, RLHF, DPO, GRPO and on-policy distillation.

Oct 15, 2026·2 min
Top themes
01NVIDIA releases Nemotron 3.5 Super: 26 AA points and 630 tokens/sOct 15, 2026·3 min02Data Rewriting Beats DP-SGD on the Privacy-Quality Trade-OffOct 15, 2026·3 min03AA-Robotics v0.1: GPT-6 Astra controls a robot arm and scores 69 out of 100Oct 14, 2026·3 min04MolLean Makes LLMs Check Their Chemistry: Wins on 12 of 12 TasksOct 14, 2026·3 min05iRTE 2026: A Look at the Real-Time Models OpenAI and Thinking Machines Are Betting OnOct 13, 2026·2 min
Sberbank Launched Free GigaPlatform AI Studio for Building AI Agents Without Code
LLM·CNews

Sberbank Launched Free GigaPlatform AI Studio for Building AI Agents Without Code

Oct 13, 2026·2 min
200 OK, but the client gets nothing: three n8n agent bugs from production
LLM·Habr AI

200 OK, but the client gets nothing: three n8n agent bugs from production

Oct 13, 2026·3 min
Video-Language Models Confuse Car Maneuvers: LoRA and Probes Fix the Error
LLM·arXiv cs.CV

Video-Language Models Confuse Car Maneuvers: LoRA and Probes Fix the Error

Oct 12, 2026·2 min
Anthropic cuts distillers' extraction of Claude's reasoning by 10–100x
LLM·Anthropic Blog

Anthropic cuts distillers' extraction of Claude's reasoning by 10–100x

Oct 12, 2026·3 min
Deep Cogito trained its Cogito v1 family on Runpod infrastructure on the road to ASI
LLM·RunPod Blog

Deep Cogito trained its Cogito v1 family on Runpod infrastructure on the road to ASI

San Francisco startup Deep Cogito trained its open Cogito v1 models on Runpod's cloud. The company's stated goal is artificial superintelligence; the lineup spans…

Oct 11, 2026·3 min
I Wrote 61 Tests for My AI Mentor. Then I Found Out You Can't Trust the Tests Either
LLM·Habr AI

I Wrote 61 Tests for My AI Mentor. Then I Found Out You Can't Trust the Tests Either

A developer built 61 tests to check an AI mentor on a coding education platform. But the test bench itself had bugs: it marked cut-off answers as passing and let…

Oct 10, 2026·3 min
LVSPM Beats VGGT at Camera Pose Estimation on 16–256 Frames and Doesn't Break Down on Large Scenes
LLM·arXiv cs.CV

LVSPM Beats VGGT at Camera Pose Estimation on 16–256 Frames and Doesn't Break Down on Large Scenes

The new LVSPM model recovers camera poses and synthesizes novel views from hundreds of uncalibrated images at once, outperforming VGGT on RealEstate10k, Co3Dv2 and DL3DV.

Oct 9, 2026·3 min
SPLIT-RL Teaches VLMs to See and Reason Separately: +1.4–6.1 pp Over GRPO
LLM·arXiv cs.CV

SPLIT-RL Teaches VLMs to See and Reason Separately: +1.4–6.1 pp Over GRPO

Oct 9, 2026·2 min
SLVR Adds 14.2 Points on BLINK Relation to Qwen2.5-VL-7B Without Text CoT
LLM·arXiv cs.CV

SLVR Adds 14.2 Points on BLINK Relation to Qwen2.5-VL-7B Without Text CoT

Oct 9, 2026·2 min
CARing teaches LLMs to predict diagnoses: R@30 hits 46.52% on MIMIC-IV
LLM·arXiv cs.LG

CARing teaches LLMs to predict diagnoses: R@30 hits 46.52% on MIMIC-IV

Oct 9, 2026·2 min
Visible CoT Does Not Universally Improve SQL and pandas Code Generation, New Study Finds
LLM·arXiv cs.LG

Visible CoT Does Not Universally Improve SQL and pandas Code Generation, New Study Finds

Oct 9, 2026·2 min
LoopOPD Teaches Looped Language Models to Improve Through Their Own Loops
LLM·arXiv cs.LG

LoopOPD Teaches Looped Language Models to Improve Through Their Own Loops

Oct 9, 2026·2 min
Lapras replaces CoT with latent reasoning: up to +10.79% F1 with 23.9x fewer tokens
LLM

Lapras replaces CoT with latent reasoning: up to +10.79% F1 with 23.9x fewer tokens

Oct 9, 2026·2 min
STI-OPD: A Probabilistic Teacher Intervention for Multi-Step Agentic Tasks
LLM

STI-OPD: A Probabilistic Teacher Intervention for Multi-Step Agentic Tasks

Oct 9, 2026·2 min
StoreBench: LLM Agents Run an Online Store and Lose to a Script — 49% vs. 97%
LLM

StoreBench: LLM Agents Run an Online Store and Lose to a Script — 49% vs. 97%

Oct 9, 2026·3 min
Midjourney Adds Thinking Mode for V8.2 and Shared Folders in Alpha
LLM·Midjourney Updates

Midjourney Adds Thinking Mode for V8.2 and Shared Folders in Alpha

Midjourney's alpha gets a thinking mode that reruns V8.2 jobs after the model checks its own output against the prompt. Shared folders add Collaborators and Viewers…

Oct 9, 2026·3 min
Midjourney Tests Thinking Mode: A Rerun (Thinking) Button for Better Text Accuracy
LLM·Midjourney Updates

Midjourney Tests Thinking Mode: A Rerun (Thinking) Button for Better Text Accuracy

Midjourney has launched Thinking Mode on its Alpha site. Rerunning an image with added "thinking" should improve prompt accuracy, typography and coherence.

Oct 9, 2026·2 min