FRI, 21 AUG 2026
Live · Daily AI brief from inside the industry
14:25:42 UTC
MODELS · 257 stories

Models.

Frontier model releases, benchmark results, and the research advancing what AI systems can do.

257
Stories in section
Time:
Sort
Anonymous lab ships free 1M-context reasoning model on OpenRouter, featured image for AI Insiders
Models / -1175 min ago

Anonymous lab ships free 1M-context reasoning model on OpenRouter

By Alessandro Benigni
Harvey Post-Trains a Chinese Open-Weight Model for Legal Work, featured image for AI Insiders
Models / -1175 min ago

Harvey Post-Trains a Chinese Open-Weight Model for Legal Work

By Alessandro Benigni
Micron commits $10 billion to a decade-long AI memory lab in Boise, featured image for AI Insiders
Models / -1175 min ago

Micron commits $10 billion to a decade-long AI memory lab in Boise

By Alessandro Benigni
Waymo Built Its Own Robotaxi Chip to Cut Nvidia Reliance, featured image for AI Insiders
Models / -1175 min ago

Waymo Built Its Own Robotaxi Chip to Cut Nvidia Reliance

By Alessandro Benigni
Anthropic Keeps Its Most Capable Model Internal, Report Shows, featured image for AI Insiders
Models / 4 hr ago

Anthropic Keeps Its Most Capable Model Internal, Report Shows

By Alessandro Benigni
DeepReinforce's Ornith-1.5 Sets Its Own Training Curriculum, featured image for AI Insiders
Models / 4 hr ago

DeepReinforce's Ornith-1.5 Sets Its Own Training Curriculum

By Alessandro Benigni
Meta's Muse Video Reaches Closed Beta With Native Audio, featured image for AI Insiders
Models / 4 hr ago

Meta's Muse Video Reaches Closed Beta With Native Audio

By Alessandro Benigni
Superwhisper's s1-mini Does One Thing: Clean Up ASR Text, featured image for AI Insiders
Models / 4 hr ago

Superwhisper's s1-mini Does One Thing: Clean Up ASR Text

By Alessandro Benigni
A Laptop-Sized Model Ties the Cloud on Quality, Not Speed, featured image for AI Insiders
Models / yesterday

A Laptop-Sized Model Ties the Cloud on Quality, Not Speed

By Alessandro Benigni
Why Thinking Machines Built Inkling Without a Vision Encoder, featured image for AI Insiders
Models / yesterday

Why Thinking Machines Built Inkling Without a Vision Encoder

By Alessandro Benigni
A Benchmark That Grades AI on Finding Out, Not Knowing, featured image for AI Insiders
Models / 2 days ago

A Benchmark That Grades AI on Finding Out, Not Knowing

By Alessandro Benigni
A Formula for How Often to Recycle Scarce Training Data, featured image for AI Insiders
Models / 2 days ago

A Formula for How Often to Recycle Scarce Training Data

By Alessandro Benigni
Magic Hour Tests Fable 5 and Sol 5.6 on Ad-Making Taste, Both Flop, featured image for AI Insiders
Models / 2 days ago

Magic Hour Tests Fable 5 and Sol 5.6 on Ad-Making Taste, Both Flop

By Alessandro Benigni
On a 24GB GPU, Qwen3.8-27B Beats Gemma 4 as Context Grows, featured image for AI Insiders
Models / 2 days ago

On a 24GB GPU, Qwen3.8-27B Beats Gemma 4 as Context Grows

By Alessandro Benigni
OrcaRouter Publishes Refusal-Stripped Qwen3.8-27B for Apple Silicon, featured image for AI Insiders
Models / 2 days ago

OrcaRouter Publishes Refusal-Stripped Qwen3.8-27B for Apple Silicon

By Alessandro Benigni
Test-Time Training Could Make Inference as Costly as Training, featured image for AI Insiders
Models / 2 days ago

Test-Time Training Could Make Inference as Costly as Training

By Alessandro Benigni
Vercel Data: Anthropic Takes 65% of Gateway Spend on 30% of Tokens, featured image for AI Insiders
Models / 2 days ago

Vercel Data: Anthropic Takes 65% of Gateway Spend on 30% of Tokens

By Alessandro Benigni
A Language Model That Only Knows What a Fifth Grader Knows, featured image for AI Insiders
Models / 3 days ago

A Language Model That Only Knows What a Fifth Grader Knows

By Alessandro Benigni
GLM-5.3 leads a cybersecurity benchmark, and Z.ai still opens the weights, featured image for AI Insiders
Models / 3 days ago

GLM-5.3 leads a cybersecurity benchmark, and Z.ai still opens the weights

By Alessandro Benigni
Google Reportedly Taps AMD to Design a CPU-Heavy TPU, featured image for AI Insiders
Models / 3 days ago

Google Reportedly Taps AMD to Design a CPU-Heavy TPU

By Alessandro Benigni
Hugging Face Data Shows Hype and Usage Split on Open Models, featured image for AI Insiders
Models / 3 days ago

Hugging Face Data Shows Hype and Usage Split on Open Models

By Alessandro Benigni
Microsoft preprint stops transformers from discarding their own state, featured image for AI Insiders
Models / 3 days ago

Microsoft preprint stops transformers from discarding their own state

By Alessandro Benigni
Qwen 3.8 27B Is Great, But Its Default Reasoning Wastes Time, featured image for AI Insiders
Models / 3 days ago

Qwen 3.8 27B Is Great, But Its Default Reasoning Wastes Time

By Alessandro Benigni
Apple Built a China-Only AI Model With Alibaba, Reuters Reports, featured image for AI Insiders
Models / 6 days ago

Apple Built a China-Only AI Model With Alibaba, Reuters Reports

By Alessandro Benigni
Gemini 3.7 Flash ships with a price cut that expires in January, featured image for AI Insiders
Models / 6 days ago

Gemini 3.7 Flash ships with a price cut that expires in January

By Alessandro Benigni
DeepSeek Prices V4-Pro at $0.87 per Million Output Tokens, featured image for AI Insiders
Models / Aug 14

DeepSeek Prices V4-Pro at $0.87 per Million Output Tokens

By Alessandro Benigni
Microsoft Ships MAI-Thinking-1, Its Own Reasoning Model for Copilot, featured image for AI Insiders
Models / Aug 14

Microsoft Ships MAI-Thinking-1, Its Own Reasoning Model for Copilot

By Alessandro Benigni
Qwen3.8's Weights Land: 2.4T Parameters, 95B Active Per Token, featured image for AI Insiders
Models / Aug 14

Qwen3.8's Weights Land: 2.4T Parameters, 95B Active Per Token

By Alessandro Benigni
xAI Releases Grok 4.6, Aimed at Long, Multi-Step Agent Work, featured image for AI Insiders
Models / Aug 14

xAI Releases Grok 4.6, Aimed at Long, Multi-Step Agent Work

By Alessandro Benigni
Gemini App Passes 1 Billion Monthly Users, Its 14th at Google, featured image for AI Insiders
Models / Aug 13

Gemini App Passes 1 Billion Monthly Users, Its 14th at Google

By Alessandro Benigni
Google Puts a Product Chief, Not a Scientist, Atop DeepMind, featured image for AI Insiders
Models / Aug 13

Google Puts a Product Chief, Not a Scientist, Atop DeepMind

By Alessandro Benigni
Microsoft's New Coding Model Beats Its Old One, Not DeepSeek, featured image for AI Insiders
Models / Aug 13

Microsoft's New Coding Model Beats Its Old One, Not DeepSeek

By Alessandro Benigni
NVIDIA fixes video world models' memory blind spot without retraining, featured image for AI Insiders
Models / Aug 13

NVIDIA fixes video world models' memory blind spot without retraining

By Alessandro Benigni
Nvidia's Nemotron 4 targets a trillion parameters, a bar China already cleared, featured image for AI Insiders
Models / Aug 13

Nvidia's Nemotron 4 targets a trillion parameters, a bar China already cleared

By Alessandro Benigni
A New Paper Asks What Two Thirds of a Transformer's Weights Buy, featured image for AI Insiders
Models / Aug 12

A New Paper Asks What Two Thirds of a Transformer's Weights Buy

By Alessandro Benigni
Claude Pushed a Riemann Bound to 67.2%. The Hypothesis Still Stands., featured image for AI Insiders
Models / Aug 12

Claude Pushed a Riemann Bound to 67.2%. The Hypothesis Still Stands.

By Alessandro Benigni
Dyna claims a scaling law: more human video, better robot performance, featured image for AI Insiders
Models / Aug 12

Dyna claims a scaling law: more human video, better robot performance

By Alessandro Benigni
Meta Opens a 30B Model Built to Live on Your Laptop, Not a Rented GPU, featured image for AI Insiders
Models / Aug 12

Meta Opens a 30B Model Built to Live on Your Laptop, Not a Rented GPU

By Alessandro Benigni
Microsoft Plans Maia 300 AI Chip Reveal in September, The Information Says, featured image for AI Insiders
Models / Aug 12

Microsoft Plans Maia 300 AI Chip Reveal in September, The Information Says

By Alessandro Benigni
OpenAI ships a hacking model and decides who may hold it, featured image for AI Insiders
Models / Aug 12

OpenAI ships a hacking model and decides who may hold it

By Alessandro Benigni
With Labs Silent on Training Specs, Guesswork Fills the Vacuum, featured image for AI Insiders
Models / Aug 12

With Labs Silent on Training Specs, Guesswork Fills the Vacuum

By Alessandro Benigni
xAI's Grok Gets a Top-Ranked Image Model, No API Yet, featured image for AI Insiders
Models / Aug 11

xAI's Grok Gets a Top-Ranked Image Model, No API Yet

By Alessandro Benigni
ByteDance Reportedly Pre-Training a 10-Trillion-Parameter Model, featured image for AI Insiders
Models / Aug 8

ByteDance Reportedly Pre-Training a 10-Trillion-Parameter Model

By Alessandro Benigni
DeepMind Open-Sources a Cyclone Model That Buys a Day of Warning, featured image for AI Insiders
Models / Aug 8

DeepMind Open-Sources a Cyclone Model That Buys a Day of Warning

By Alessandro Benigni
Meta Preprint Finds a Training Recipe Needing 5% of the Compute, featured image for AI Insiders
Models / Aug 8

Meta Preprint Finds a Training Recipe Needing 5% of the Compute

By Alessandro Benigni
OpenAI puts free ChatGPT users on Luna, keeps Sol for paying tiers, featured image for AI Insiders
Models / Aug 8

OpenAI puts free ChatGPT users on Luna, keeps Sol for paying tiers

By Alessandro Benigni
ByteDance's Seed Lab Ships an AI That Watches While It Talks, featured image for AI Insiders
Models / Aug 7

ByteDance's Seed Lab Ships an AI That Watches While It Talks

By Alessandro Benigni
Hassabis Moves Upstairs, Jeff Dean Walks as Alphabet Sheds 5%, featured image for AI Insiders
Models / Aug 7

Hassabis Moves Upstairs, Jeff Dean Walks as Alphabet Sheds 5%

By Alessandro Benigni
Qwen3.8 Max Matches Claude Opus 4.8, Then the Bill Arrives, featured image for AI Insiders
Models / Aug 7

Qwen3.8 Max Matches Claude Opus 4.8, Then the Bill Arrives

By Alessandro Benigni
Xiaomi Open-Sources a Robot Foundation Model, Not Just a Demo, featured image for AI Insiders
Models / Aug 7

Xiaomi Open-Sources a Robot Foundation Model, Not Just a Demo

By Alessandro Benigni
Black Forest Labs opens FLUX 3 Video, cites lead over rivals, featured image for AI Insiders
Models / Aug 6

Black Forest Labs opens FLUX 3 Video, cites lead over rivals

By Alessandro Benigni
DiffusionGemma's own paper says 1,500 tokens per second, not 1,000, featured image for AI Insiders
Models / Aug 6

DiffusionGemma's own paper says 1,500 tokens per second, not 1,000

By Alessandro Benigni
Liquid AI's 2.6B On-Device Model Bets Agents Beat Chat, featured image for AI Insiders
Models / Aug 6

Liquid AI's 2.6B On-Device Model Bets Agents Beat Chat

By Alessandro Benigni
Mistral Puts the Moderation Policy in the Prompt, Not the Weights, featured image for AI Insiders
Models / Aug 6

Mistral Puts the Moderation Policy in the Prompt, Not the Weights

By Alessandro Benigni
NVIDIA Clears Alpamayo 2 Super for Commercial Robotaxi Use, featured image for AI Insiders
Models / Aug 6

NVIDIA Clears Alpamayo 2 Super for Commercial Robotaxi Use

By Alessandro Benigni
NVIDIA open-sources an 11B voice model with live tool calling, featured image for AI Insiders
Models / Aug 6

NVIDIA open-sources an 11B voice model with live tool calling

By Alessandro Benigni
A spy-game reward teaches AI what no answer key can grade, featured image for AI Insiders
Models / Aug 5

A spy-game reward teaches AI what no answer key can grade

By Alessandro Benigni
Independent test puts a real price on DeepSeek's cheap model, featured image for AI Insiders
Models / Aug 5

Independent test puts a real price on DeepSeek's cheap model

By Alessandro Benigni
Mind Lab says patching GLM-5.2 with 4B params beats the base model, featured image for AI Insiders
Models / Aug 5

Mind Lab says patching GLM-5.2 with 4B params beats the base model

By Alessandro Benigni
OpenAI's Same Token Price Hides a 2.25x Usage Jump, featured image for AI Insiders
Models / Aug 5

OpenAI's Same Token Price Hides a 2.25x Usage Jump

By Alessandro Benigni
DeepSeek ships production V4-Flash-0731, claims wins over its own Pro model, featured image for AI Insiders
Models / Aug 4

DeepSeek ships production V4-Flash-0731, claims wins over its own Pro model

By Alessandro Benigni
Microsoft Tests Native Full-Duplex Voice Model MAI Realtime, featured image for AI Insiders
Models / Aug 4

Microsoft Tests Native Full-Duplex Voice Model MAI Realtime

By Alessandro Benigni
MiniMax's H3 weights are out, and open has conditions, featured image for AI Insiders
Models / Aug 4

MiniMax's H3 weights are out, and open has conditions

By Alessandro Benigni
OpenAI's Unreleased Astra Model Cracks Ten Open Math Problems, featured image for AI Insiders
Models / Aug 4

OpenAI's Unreleased Astra Model Cracks Ten Open Math Problems

By Alessandro Benigni
Qwen3.8-Max Ships With a 16-Day Autonomous Coding Marathon, featured image for AI Insiders
Models / Aug 4

Qwen3.8-Max Ships With a 16-Day Autonomous Coding Marathon

By Alessandro Benigni
Arcee's Trinity Mini Learns to Search Evidence, Not Do Science, featured image for AI Insiders
Models / Aug 1

Arcee's Trinity Mini Learns to Search Evidence, Not Do Science

By Alessandro Benigni
DeepSeek's Cheap Model Nearly Matches GPT-5.6 Luna, Cheaper Still, featured image for AI Insiders
Models / Aug 1

DeepSeek's Cheap Model Nearly Matches GPT-5.6 Luna, Cheaper Still

By Alessandro Benigni
Google DeepMind launches Gemini Robotics ER 2 for robots, featured image for AI Insiders
Models / Aug 1

Google DeepMind launches Gemini Robotics ER 2 for robots

By Alessandro Benigni
MiniMax unveils H3, an open video model with no weights link, featured image for AI Insiders
Models / Aug 1

MiniMax unveils H3, an open video model with no weights link

By Alessandro Benigni
Open-Weight Models Match GPT-5.6 Sol on Clinical Benchmark, featured image for AI Insiders
Models / Aug 1

Open-Weight Models Match GPT-5.6 Sol on Clinical Benchmark

By Alessandro Benigni
OpenAI cuts GPT-5.6 Luna price 80 percent, Terra 20 percent, featured image for AI Insiders
Models / Aug 1

OpenAI cuts GPT-5.6 Luna price 80 percent, Terra 20 percent

By Alessandro Benigni
Thinking Machines shrinks its flagship to a quarter of the compute, featured image for AI Insiders
Models / Aug 1

Thinking Machines shrinks its flagship to a quarter of the compute

By Alessandro Benigni
A DIY GPT-2 Beat OpenAI's Weights on Loss, Lost on Instructions, featured image for AI Insiders
Models / Jul 31

A DIY GPT-2 Beat OpenAI's Weights on Loss, Lost on Instructions

By Alessandro Benigni
Claude Opus 5 wins the vending machine benchmark by lying to suppliers, featured image for AI Insiders
Models / Jul 31

Claude Opus 5 wins the vending machine benchmark by lying to suppliers

By Alessandro Benigni
DeepMind Finds the Prompt Isn't Just Text, It's the Image, featured image for AI Insiders
Models / Jul 31

DeepMind Finds the Prompt Isn't Just Text, It's the Image

By Alessandro Benigni
EschaLabs Squeezes a 35B MoE Model Onto One Consumer GPU, featured image for AI Insiders
Models / Jul 31

EschaLabs Squeezes a 35B MoE Model Onto One Consumer GPU

By Alessandro Benigni
Liquid AI Ships Encoders That Skip the GPU for Long Documents, featured image for AI Insiders
Models / Jul 31

Liquid AI Ships Encoders That Skip the GPU for Long Documents

By Alessandro Benigni
Nvidia Method Shrinks AI Video Sampling to Four or Eight Steps, featured image for AI Insiders
Models / Jul 31

Nvidia Method Shrinks AI Video Sampling to Four or Eight Steps

By Alessandro Benigni
OpenAI says GPT-5.6 Sol optimized its own inference stack, featured image for AI Insiders
Models / Jul 31

OpenAI says GPT-5.6 Sol optimized its own inference stack

By Alessandro Benigni
OpenAI Says Its Own API Settings Triple ARC-AGI-3 Scores, featured image for AI Insiders
Models / Jul 31

OpenAI Says Its Own API Settings Triple ARC-AGI-3 Scores

By Alessandro Benigni
xAI's Grok Voice Think Fast 2.0 lands at $0.08 a minute, featured image for AI Insiders
Models / Jul 31

xAI's Grok Voice Think Fast 2.0 lands at $0.08 a minute

By Alessandro Benigni
Amazon Reportedly Plans to Cut Nova Down to One Frontier Model, featured image for AI Insiders
Models / Jul 30

Amazon Reportedly Plans to Cut Nova Down to One Frontier Model

By Alessandro Benigni
Anthropic's Claude finds faster attacks on HAWK and reduced AES, featured image for AI Insiders
Models / Jul 30

Anthropic's Claude finds faster attacks on HAWK and reduced AES

By Alessandro Benigni
Bagel Labs releases WorldDiT, a compact world model for robots, featured image for AI Insiders
Models / Jul 30

Bagel Labs releases WorldDiT, a compact world model for robots

By Alessandro Benigni
Fish Audio ships S2.1 Pro, a 90ms voice model across 83 languages, featured image for AI Insiders
Models / Jul 30

Fish Audio ships S2.1 Pro, a 90ms voice model across 83 languages

By Alessandro Benigni
Kimi K3's Architecture Is Old News, Just Much Bigger, featured image for AI Insiders
Models / Jul 30

Kimi K3's Architecture Is Old News, Just Much Bigger

By Alessandro Benigni
Microsoft's Mage family splits into a video reader and an image maker, featured image for AI Insiders
Models / Jul 30

Microsoft's Mage family splits into a video reader and an image maker

By Alessandro Benigni
OpenAI's New Transcriber Improves, Still Trails Three Rivals, featured image for AI Insiders
Models / Jul 30

OpenAI's New Transcriber Improves, Still Trails Three Rivals

By Alessandro Benigni
A Baseten Engineer Traces Why Kimi K3 Beats Scale Alone, featured image for AI Insiders
Models / Jul 29

A Baseten Engineer Traces Why Kimi K3 Beats Scale Alone

By Alessandro Benigni
InclusionAI's LLaDA2.2 pushes diffusion language models past 100B, featured image for AI Insiders
Models / Jul 29

InclusionAI's LLaDA2.2 pushes diffusion language models past 100B

By Alessandro Benigni
Looped transformers decouple parameter count from capability, featured image for AI Insiders
Models / Jul 29

Looped transformers decouple parameter count from capability

By Alessandro Benigni
Microsoft Bets a Purpose-Built Model Beats General AI on Vuln Hunting, featured image for AI Insiders
Models / Jul 29

Microsoft Bets a Purpose-Built Model Beats General AI on Vuln Hunting

By Alessandro Benigni
Moonshot Ships Kimi K3 Weights, Then Opens the Machinery Behind Them, featured image for AI Insiders
Models / Jul 29

Moonshot Ships Kimi K3 Weights, Then Opens the Machinery Behind Them

By Alessandro Benigni
A Legal AI Vendor Built the Benchmark Judging Its Own Product, featured image for AI Insiders
Models / Jul 28

A Legal AI Vendor Built the Benchmark Judging Its Own Product

By Alessandro Benigni
Anthropic's Opus 5 nears its top model's power at half the price, featured image for AI Insiders
Models / Jul 28

Anthropic's Opus 5 nears its top model's power at half the price

By Alessandro Benigni
Benchmark Test Pits a 3B Model Against Poolside's 118B Coding MoE, featured image for AI Insiders
Models / Jul 28

Benchmark Test Pits a 3B Model Against Poolside's 118B Coding MoE

By Alessandro Benigni
Celeris Labs Claims GPT-5-Level Speed at 157ms With Diffusion, featured image for AI Insiders
Models / Jul 28

Celeris Labs Claims GPT-5-Level Speed at 157ms With Diffusion

By Alessandro Benigni
NVIDIA's SANA-Video 2.0 renders 720p clips on one GPU, featured image for AI Insiders
Models / Jul 28

NVIDIA's SANA-Video 2.0 renders 720p clips on one GPU

By Alessandro Benigni
AMD and Cerebras Split Inference in Two to Cut Latency, featured image for AI Insiders
Models / Jul 25

AMD and Cerebras Split Inference in Two to Cut Latency

By Alessandro Benigni
Black Forest Labs Reframes Itself as a World-Model Company, featured image for AI Insiders
Models / Jul 25

Black Forest Labs Reframes Itself as a World-Model Company

By Alessandro Benigni
DeepSeek's Huawei Training Numbers Arrive. So Do the Doubts, featured image for AI Insiders
Models / Jul 25

DeepSeek's Huawei Training Numbers Arrive. So Do the Doubts

By Alessandro Benigni
Etched Says New Chip Design Runs Math Blocks at Half Voltage, featured image for AI Insiders
Models / Jul 25

Etched Says New Chip Design Runs Math Blocks at Half Voltage

By Alessandro Benigni
Kimi K3's Real Edge Is What It Does Inside Its Own Head, featured image for AI Insiders
Models / Jul 25

Kimi K3's Real Edge Is What It Does Inside Its Own Head

By Alessandro Benigni
Microsoft Ships MAI-Image-2.5-Pro and a Faster, Cheaper Voice Model, featured image for AI Insiders
Models / Jul 25

Microsoft Ships MAI-Image-2.5-Pro and a Faster, Cheaper Voice Model

By Alessandro Benigni
DOE and Arcee AI open GS1, a public open-weight science model, featured image for AI Insiders
Models / Jul 24

DOE and Arcee AI open GS1, a public open-weight science model

By Alessandro Benigni
Google Splits Gemini Flash Into Three Models for Speed, Cost, Security, featured image for AI Insiders
Models / Jul 23

Google Splits Gemini Flash Into Three Models for Speed, Cost, Security

By Alessandro Benigni
Meta's SAM 3 and DINOv3 Now Run Inside a DOE National Lab Pipeline, featured image for AI Insiders
Models / Jul 23

Meta's SAM 3 and DINOv3 Now Run Inside a DOE National Lab Pipeline

By Alessandro Benigni
Microsoft Ships Mage, a 4B Image Model Line Built for Labs, Not Products, featured image for AI Insiders
Models / Jul 23

Microsoft Ships Mage, a 4B Image Model Line Built for Labs, Not Products

By Alessandro Benigni
Poolside Ships Laguna S 2.1, a 118B Coding Model at 8B Cost, featured image for AI Insiders
Models / Jul 23

Poolside Ships Laguna S 2.1, a 118B Coding Model at 8B Cost

By Alessandro Benigni
Qwen's new image model targets slide decks, not just pretty pictures, featured image for AI Insiders
Models / Jul 23

Qwen's new image model targets slide decks, not just pretty pictures

By Alessandro Benigni
AMD's Helios rack system lands Microsoft, challenges Nvidia grip, featured image for AI Insiders
Models / Jul 22

AMD's Helios rack system lands Microsoft, challenges Nvidia grip

By Alessandro Benigni
Google Reportedly Building New Chip to Cut AI Power Costs, featured image for AI Insiders
Models / Jul 22

Google Reportedly Building New Chip to Cut AI Power Costs

By Alessandro Benigni
NVIDIA shrinks Cosmos 3 to run its world model on the robot itself, featured image for AI Insiders
Models / Jul 22

NVIDIA shrinks Cosmos 3 to run its world model on the robot itself

By Alessandro Benigni
Xiaomi's New Robot Model Trains Mostly on Human Video, Not Robots, featured image for AI Insiders
Models / Jul 22

Xiaomi's New Robot Model Trains Mostly on Human Video, Not Robots

By Alessandro Benigni
Alibaba's Qwen3.8 Joins the 2-Trillion-Parameter Open Weight Sprint. Featured image for AI Insiders.
Models / Jul 21

Alibaba's Qwen3.8 Joins the 2-Trillion-Parameter Open Weight Sprint

By Alessandro Benigni
Claude Fable 5 joins Max, Team Premium plan limits July 20. Featured image for AI Insiders.
Models / Jul 21

Claude Fable 5 joins Max, Team Premium plan limits July 20

By Alessandro Benigni
Sakana AI trains neural nets without breaking a core brain rule. Featured image for AI Insiders.
Models / Jul 21

Sakana AI trains neural nets without breaking a core brain rule

By Alessandro Benigni
Alphabet Stock Drops 4% on Report Gemini 3.5 Pro Is Delayed. Featured image for AI Insiders.
Models / Jul 18

Alphabet Stock Drops 4% on Report Gemini 3.5 Pro Is Delayed

By Alessandro Benigni
Moonshot's Kimi K3 becomes the largest open-weight model yet. Featured image for AI Insiders.
Models / Jul 18

Moonshot's Kimi K3 becomes the largest open-weight model yet

By Alessandro Benigni
NVIDIA Nemotron 3 Embed Tops RTEB, Ships as Open Weights. Featured image for AI Insiders.
Models / Jul 18

NVIDIA Nemotron 3 Embed Tops RTEB, Ships as Open Weights

By Alessandro Benigni
Nvidia's Cosmos 3 Edge lands as Fujitsu, Hitachi join Japan push — featured image for AI Insiders
Models / Jul 16

Nvidia's Cosmos 3 Edge lands as Fujitsu, Hitachi join Japan push

By Alessandro Benigni
OpenAI built an AI that breaks its own models to make them safer — featured image for AI Insiders
Models / Jul 16

OpenAI built an AI that breaks its own models to make them safer

By Alessandro Benigni
GPT-5.6 Sol Jumps 18 Spots to Top a Web Design Leaderboard — featured image for AI Insiders
Models / Jul 16

GPT-5.6 Sol Jumps 18 Spots to Top a Web Design Leaderboard

By Alessandro Benigni
Thinking Machines Ships Its First Open Model, a 975B MoE — featured image for AI Insiders
Models / Jul 16

Thinking Machines Ships Its First Open Model, a 975B MoE

By Alessandro Benigni
Gemma 4 gets a Pixel 10 chip variant that also drives the phone — featured image for AI Insiders
Models / Jul 16

Gemma 4 gets a Pixel 10 chip variant that also drives the phone

By Alessandro Benigni
Open weights now carry most production AI traffic, report finds — featured image for AI Insiders
Models / Jul 16

Open weights now carry most production AI traffic, report finds

By Alessandro Benigni
PrismML says its 27B model now fits on a phone, at 3.9GB — featured image for AI Insiders
Models / Jul 16

PrismML says its 27B model now fits on a phone, at 3.9GB

By Alessandro Benigni
DeepMind Repurposes a Video Generator as a Universal Vision Model — featured image for AI Insiders
Models / Jul 15

DeepMind Repurposes a Video Generator as a Universal Vision Model

By Alessandro Benigni
Monzo Co-Founder Tom Blomfield Leaves YC to Join Anthropic's Compute Team — featured image for AI Insiders
Models / Jul 15

Monzo Co-Founder Tom Blomfield Leaves YC to Join Anthropic's Compute Team

By Alessandro Benigni
Researchers Say AI's Next Scaling Axis Is Better Verification — featured image for AI Insiders
Models / Jul 13

Researchers Say AI's Next Scaling Axis Is Better Verification

By Alessandro Benigni
Anthropic Extends Fable 5 Access Again, Now Through July 19 — featured image for AI Insiders
Models / Jul 13

Anthropic Extends Fable 5 Access Again, Now Through July 19

By Alessandro Benigni
GPT-5.6 Sol Wins an ARC-AGI-3 Game, Yet Still Fails Most of the Test — featured image for AI Insiders
Models / Jul 11

GPT-5.6 Sol Wins an ARC-AGI-3 Game, Yet Still Fails Most of the Test

By Alessandro Benigni
Meta Opens Its Model API as Muse Spark Gets a 1.1 Upgrade — featured image for AI Insiders
Models / Jul 11

Meta Opens Its Model API as Muse Spark Gets a 1.1 Upgrade

By Alessandro Benigni
Meta's Custom AI Chips Head Into Production This September — featured image for AI Insiders
Models / Jul 11

Meta's Custom AI Chips Head Into Production This September

By Alessandro Benigni
Nvidia's Flex-Forcing Merges Two Video Generation Modes Into One Model — featured image for AI Insiders
Models / Jul 11

Nvidia's Flex-Forcing Merges Two Video Generation Modes Into One Model

By Alessandro Benigni
OpenAI ships GPT-5.6 in three tiers, Sol, Terra, Luna, after a gate — featured image for AI Insiders
Models / Jul 11

OpenAI ships GPT-5.6 in three tiers, Sol, Terra, Luna, after a gate

By Alessandro Benigni
Z.ai Ditches GRPO's Grouped Sampling to Fix Async RL Instability — featured image for AI Insiders
Models / Jul 11

Z.ai Ditches GRPO's Grouped Sampling to Fix Async RL Instability

By Alessandro Benigni
Mistral's Robostral Navigate Steers Robots With One Camera — featured image for AI Insiders
Models / Jul 9

Mistral's Robostral Navigate Steers Robots With One Camera

By Alessandro Benigni
Cognition's SWE-1.7 Chases Frontier Coding at Lower Cost — featured image for AI Insiders
Models / Jul 9

Cognition's SWE-1.7 Chases Frontier Coding at Lower Cost

By Alessandro Benigni
OpenAI Finds Roughly 30% of SWE-Bench Pro Tasks Are Broken — featured image for AI Insiders
Models / Jul 9

OpenAI Finds Roughly 30% of SWE-Bench Pro Tasks Are Broken

By Alessandro Benigni
ByteDance's Seedream 5.0 Pro Bets on Editing, Not One-Shot Images — featured image for AI Insiders
Models / Jul 9

ByteDance's Seedream 5.0 Pro Bets on Editing, Not One-Shot Images

By Alessandro Benigni
OpenAI Launches GPT-Live, a Voice Model That Talks and Listens at Once — featured image for AI Insiders
Models / Jul 9

OpenAI Launches GPT-Live, a Voice Model That Talks and Listens at Once

By Alessandro Benigni
SpaceXAI to Launch Grok 4.5 Thursday, Musk Claims Opus-Class Power — featured image for AI Insiders
Models / Jul 8

SpaceXAI to Launch Grok 4.5 Thursday, Musk Claims Opus-Class Power

By Alessandro Benigni
Anthropic Offers Free Claude Fable 5 Access Through July 12 — featured image for AI Insiders
Models / Jul 8

Anthropic Offers Free Claude Fable 5 Access Through July 12

By Alessandro Benigni
Liquid AI's Antidoom Nearly Eliminates Reasoning-Model Repetition Loops — featured image for AI Insiders
Models / Jul 8

Liquid AI's Antidoom Nearly Eliminates Reasoning-Model Repetition Loops

By Alessandro Benigni
Meta's Muse Image skips the API, lands in WhatsApp and IG instead — featured image for AI Insiders
Models / Jul 8

Meta's Muse Image skips the API, lands in WhatsApp and IG instead

By Alessandro Benigni
Gemma 4's technical report bets on ditching AI encoders — featured image for AI Insiders
Models / Jul 8

Gemma 4's technical report bets on ditching AI encoders

By Alessandro Benigni
Microsoft Shifts Excel and Outlook AI to Its Own MAI Models — featured image for AI Insiders
Models / Jul 8

Microsoft Shifts Excel and Outlook AI to Its Own MAI Models

By Alessandro Benigni
OpenAI Sets GPT-5.6 Public Launch for Thursday, Undercuts 5.5 on Price — featured image for AI Insiders
Models / Jul 8

OpenAI Sets GPT-5.6 Public Launch for Thursday, Undercuts 5.5 on Price

By Alessandro Benigni
Anthropic Finds a Workspace for Deliberate Thought in Claude — featured image for AI Insiders
Models / Jul 8

Anthropic Finds a Workspace for Deliberate Thought in Claude

By Alessandro Benigni
Broadcom Locks In Apple Silicon Deal Through 2031 — featured image for AI Insiders
Models / Jul 8

Broadcom Locks In Apple Silicon Deal Through 2031

By Alessandro Benigni
Tencent Ships Hy3, a 295B Open Model, Free Through July 21 — featured image for AI Insiders
Models / Jul 8

Tencent Ships Hy3, a 295B Open Model, Free Through July 21

By Alessandro Benigni
ByteDance targets July 9 for Seedance 2.5's 3-minute leap. Featured image for AI Insiders
Models / Jul 7

ByteDance targets July 9 for Seedance 2.5's 3-minute leap

By Alessandro Benigni
OpenAI Splits GPT-5.6 Into Three Tiers, Adds Reasoning Slider. Featured image for AI Insiders
Models / Jul 7

OpenAI Splits GPT-5.6 Into Three Tiers, Adds Reasoning Slider

By Alessandro Benigni
ByteDance Seed's Model Card Puts Evaluation Design Before Benchmarks — featured image for AI Insiders
Models / Jul 3

ByteDance Seed's Model Card Puts Evaluation Design Before Benchmarks

By Alessandro Benigni
Poolside's Laguna XS 2.1 lifts SWE-bench score, loosens its license — featured image for AI Insiders
Models / Jul 3

Poolside's Laguna XS 2.1 lifts SWE-bench score, loosens its license

By Alessandro Benigni
Apple Recycles the Tokens Diffusion Language Models Throw Away — featured image for AI Insiders
Models / Jul 3

Apple Recycles the Tokens Diffusion Language Models Throw Away

By Alessandro Benigni
Meta's unreleased Watermelon model reportedly closes gap with GPT-5.5 — featured image for AI Insiders
Models / Jul 3

Meta's unreleased Watermelon model reportedly closes gap with GPT-5.5

By Alessandro Benigni
Google tests new Gemini Flash build on LM Arena — featured image for AI Insiders
Models / Jul 2

Google tests new Gemini Flash build on LM Arena

By Alessandro Benigni
GPT-5.5 Pro and Claude Opus 4.7 paired up to crack open math problems — featured image for AI Insiders
Models / Jul 1

GPT-5.5 Pro and Claude Opus 4.7 paired up to crack open math problems

By Alessandro Benigni
Meituan's LongCat-2.0 outs itself as OpenRouter's stealth hit — featured image for AI Insiders
Models / Jul 1

Meituan's LongCat-2.0 outs itself as OpenRouter's stealth hit

By Alessandro Benigni
Anthropic prices Claude Sonnet 5 to undercut its own Opus tier — featured image for AI Insiders
Models / Jul 1

Anthropic prices Claude Sonnet 5 to undercut its own Opus tier

By Alessandro Benigni
AI2's DiScoFormer cuts density error 37x over KDE at 100 dimensions — featured image for AI Insiders
Models / Jul 1

AI2's DiScoFormer cuts density error 37x over KDE at 100 dimensions

By Alessandro Benigni
Google Cloud will sell SandboxAQ's science models alongside Gemini — featured image for AI Insiders
Models / Jul 1

Google Cloud will sell SandboxAQ's science models alongside Gemini

By Alessandro Benigni
Google opens Gemini's personalized image generation to all US free users — featured image for AI Insiders
Models / Jul 1

Google opens Gemini's personalized image generation to all US free users

By Alessandro Benigni
Sakana AI ships Fugu Ultra at 93.2 on LiveCodeBench after losing Claude — featured image for AI Insiders
Models / Jul 1

Sakana AI ships Fugu Ultra at 93.2 on LiveCodeBench after losing Claude

By Alessandro Benigni
Google retrofits multi-token prediction onto frozen Gemini Nano — featured image for AI Insiders
Models / Jun 30

Google retrofits multi-token prediction onto frozen Gemini Nano

By Alessandro Benigni
Grok 4.5 enters private beta at SpaceX and Tesla on a 1.5T-parameter base — featured image for AI Insiders
Models / Jun 30

Grok 4.5 enters private beta at SpaceX and Tesla on a 1.5T-parameter base

By Alessandro Benigni
OpenAI previews GPT-5.6 under U.S. government watch — featured image for AI Insiders
Models / Jun 30

OpenAI previews GPT-5.6 under U.S. government watch

By Alessandro Benigni
Reward Models Are Too Sensitive, and That Is Why Your RLHF Breaks — featured image for AI Insiders
Models / Jun 30

Reward Models Are Too Sensitive, and That Is Why Your RLHF Breaks

By Alessandro Benigni
DeepReinforce ships Ornith-1.0, a coding model that writes its own RL scaffolds — featured image for AI Insiders
Models / Jun 27

DeepReinforce ships Ornith-1.0, a coding model that writes its own RL scaffolds

By Alessandro Benigni
Goodfire removed a model's German fluency by tuning 4 tokens — featured image for AI Insiders
Models / Jun 27

Goodfire removed a model's German fluency by tuning 4 tokens

By Alessandro Benigni
Liquid AI shrinks its non-transformer model to 230M parameters — featured image for AI Insiders
Models / Jun 27

Liquid AI shrinks its non-transformer model to 230M parameters

By Alessandro Benigni
GLM-5.2 is the first open model that holds up in a coding agent harness — featured image for AI Insiders
Models / Jun 26

GLM-5.2 is the first open model that holds up in a coding agent harness

By Alessandro Benigni
OpenAI and Broadcom build Jalapeño, a custom LLM inference chip — featured image for AI Insiders
Models / Jun 26

OpenAI and Broadcom build Jalapeño, a custom LLM inference chip

By Alessandro Benigni
OpenAI Quietly Upgrades GPT-5.5 Instant Inside ChatGPT — featured image for AI Insiders
Models / Jun 26

OpenAI Quietly Upgrades GPT-5.5 Instant Inside ChatGPT

By Alessandro Benigni
Baidu's Unlimited OCR parses dozens of pages in a single forward pass — featured image for AI Insiders
Models / Jun 25

Baidu's Unlimited OCR parses dozens of pages in a single forward pass

By Alessandro Benigni
ByteDance's Seedance 2.5 Generates 30-Second 4K Video From One Prompt — featured image for AI Insiders
Models / Jun 25

ByteDance's Seedance 2.5 Generates 30-Second 4K Video From One Prompt

By Alessandro Benigni
Krea releases open-weight image models built to escape default aesthetics — featured image for AI Insiders
Models / Jun 25

Krea releases open-weight image models built to escape default aesthetics

By Alessandro Benigni
A 0.22B inpainting model that matches an 11.9B generalist — featured image for AI Insiders
Models / Jun 24

A 0.22B inpainting model that matches an 11.9B generalist

By Alessandro Benigni
Alibaba's HappyHorse 1.1 passes Sora in video rankings — featured image for AI Insiders
Models / Jun 24

Alibaba's HappyHorse 1.1 passes Sora in video rankings

By Alessandro Benigni
The transformer monoculture is over. Here is what replaced it. — featured image for AI Insiders
Models / Jun 22

The transformer monoculture is over. Here is what replaced it.

By Alessandro Benigni
Diffusion LLMs Are Not an Interpretability Dead End — featured image for AI Insiders
Models / Jun 22

Diffusion LLMs Are Not an Interpretability Dead End

By Alessandro Benigni
Mercury 2 hits 1,000 tokens per second. Here is what that buys you. — featured image for AI Insiders
Models / Jun 22

Mercury 2 hits 1,000 tokens per second. Here is what that buys you.

By Alessandro Benigni
NVIDIA's ZPPO Fixes the Hardest-Question Problem in RL Post-Training — featured image for AI Insiders
Models / Jun 20

NVIDIA's ZPPO Fixes the Hardest-Question Problem in RL Post-Training

By Alessandro Benigni
OpenAI eyes GPT-5.6 launch next week with 1.5M token context — featured image for AI Insiders
Models / Jun 20

OpenAI eyes GPT-5.6 launch next week with 1.5M token context

By Alessandro Benigni
OpenAI shows RL on beneficial traits generalizes across dozens of benchmarks — featured image for AI Insiders
Models / Jun 20

OpenAI shows RL on beneficial traits generalizes across dozens of benchmarks

By Alessandro Benigni
Ai2 ships MolmoMotion to close robotics' language gap — featured image for AI Insiders
Models / Jun 19

Ai2 ships MolmoMotion to close robotics' language gap

By Alessandro Benigni
Cursor is building a 1.5-trillion-parameter model from scratch — featured image for AI Insiders
Models / Jun 19

Cursor is building a 1.5-trillion-parameter model from scratch

By Alessandro Benigni
Kimi K2.7 Code ran 16x cheaper than Claude Fable 5 in a landing-page test — featured image for AI Insiders
Models / Jun 19

Kimi K2.7 Code ran 16x cheaper than Claude Fable 5 in a landing-page test

By Alessandro Benigni
Mistral Previews a Large Sparse MoE This Summer, Opens July Access — featured image for AI Insiders
Models / Jun 19

Mistral Previews a Large Sparse MoE This Summer, Opens July Access

By Alessandro Benigni
OpenAI's new life-science benchmark hits a 36% ceiling — featured image for AI Insiders
Models / Jun 19

OpenAI's new life-science benchmark hits a 36% ceiling

By Alessandro Benigni
Microsoft Extends Phi Silica to NVIDIA GPUs, Testing NPU Limits — featured image for AI Insiders
Models / Jun 18

Microsoft Extends Phi Silica to NVIDIA GPUs, Testing NPU Limits

By Alessandro Benigni
OpenAI's GPT-Bidi-1 aims to fix voice mode's turn-taking problem — featured image for AI Insiders
Models / Jun 18

OpenAI's GPT-Bidi-1 aims to fix voice mode's turn-taking problem

By Alessandro Benigni
Qwen-RobotWorld makes plain language the control layer for robots — featured image for AI Insiders
Models / Jun 18

Qwen-RobotWorld makes plain language the control layer for robots

By Alessandro Benigni
Weibo's 3B model matches flagships on LeetCode, falls short on science — featured image for AI Insiders
Models / Jun 18

Weibo's 3B model matches flagships on LeetCode, falls short on science

By Alessandro Benigni
Z.ai ships GLM-5.2 with a 1M-token context for full codebases — featured image for AI Insiders
Models / Jun 18

Z.ai ships GLM-5.2 with a 1M-token context for full codebases

By Alessandro Benigni
Moonshot AI ships Kimi K2.7 Code, a 1T-parameter MoE drop-in for OpenAI APIs — featured image for AI Insiders
Models / Jun 16

Moonshot AI ships Kimi K2.7 Code, a 1T-parameter MoE drop-in for OpenAI APIs

By Alessandro Benigni
Z.ai's GLM-5.2 ships with MIT license and 1M-context coding support — featured image for AI Insiders
Models / Jun 16

Z.ai's GLM-5.2 ships with MIT license and 1M-context coding support

By Alessandro Benigni
Google ships DiffusionGemma, a 26B open model that generates text 4x faster — featured image for AI Insiders
Models / Jun 12

Google ships DiffusionGemma, a 26B open model that generates text 4x faster

By Alessandro Benigni
The Fable 5 leak's real story is 120,000 characters — featured image for AI Insiders
Models / Jun 12

The Fable 5 leak's real story is 120,000 characters

By Alessandro Benigni
Anthropic ships Mythos-class capability to the public via Claude Fable 5 — featured image for AI Insiders
Models / Jun 11

Anthropic ships Mythos-class capability to the public via Claude Fable 5

By Alessandro Benigni
Xiaomi hits 1,000 tokens/sec on a trillion-param model using commodity GPUs — featured image for AI Insiders
Models / Jun 10

Xiaomi hits 1,000 tokens/sec on a trillion-param model using commodity GPUs

By Alessandro Benigni
Apple pays Google $1B/yr to admit it lost the AI race — featured image for AI Insiders
Models / Jun 9

Apple pays Google $1B/yr to admit it lost the AI race

By Alessandro Benigni
Claude matches ChemDraw on NMR, then does what ChemDraw cannot — featured image for AI Insiders
Models / Jun 9

Claude matches ChemDraw on NMR, then does what ChemDraw cannot

By Alessandro Benigni
Google ships Gemma 4 QAT models that fit in 1GB on a phone — featured image for AI Insiders
Models / Jun 9

Google ships Gemma 4 QAT models that fit in 1GB on a phone

By Alessandro Benigni
Anthropic's Oceanus red team leaked to Chinese proxy within hours — featured image for AI Insiders
Models / Jun 6

Anthropic's Oceanus red team leaked to Chinese proxy within hours

By Alessandro Benigni
NVIDIA ships a unified safety model with auditable reasoning — featured image for AI Insiders
Models / Jun 6

NVIDIA ships a unified safety model with auditable reasoning

By Alessandro Benigni
Google's Sleep+Dreaming turns idle time into a training loop — featured image for AI Insiders
Models / Jun 5

Google's Sleep+Dreaming turns idle time into a training loop

By Alessandro Benigni
Meta's Muse Spark Has No API Launch Date — featured image for AI Insiders
Models / Jun 5

Meta's Muse Spark Has No API Launch Date

By Alessandro Benigni
Microsoft's MAI bets enterprises will tune their own models — featured image for AI Insiders
Models / Jun 4

Microsoft's MAI bets enterprises will tune their own models

By Alessandro Benigni
MiniMax M3 ships API-first with weights still pending — featured image for AI Insiders
Models / Jun 4

MiniMax M3 ships API-first with weights still pending

By Alessandro Benigni
Alibaba's Qwen3.7-Plus aims to unify GUI and CLI in one agent loop — featured image for AI Insiders
Models / Jun 3

Alibaba's Qwen3.7-Plus aims to unify GUI and CLI in one agent loop

By Alessandro Benigni
JetBrains ships Mellum 2, a 12B MoE model built for IDE-native coding — featured image for AI Insiders
Models / Jun 3

JetBrains ships Mellum 2, a 12B MoE model built for IDE-native coding

By Alessandro Benigni
NVIDIA Ships Cosmos 3, an Open Foundation Model That Outputs Robot Actions — featured image for AI Insiders
Models / Jun 3

NVIDIA Ships Cosmos 3, an Open Foundation Model That Outputs Robot Actions

By Alessandro Benigni
NVIDIA's Nemotron 3 Ultra is the strongest US open-weights model — featured image for AI Insiders
Models / Jun 3

NVIDIA's Nemotron 3 Ultra is the strongest US open-weights model

By Alessandro Benigni
MiniMax M3 beats Opus 4.7 on BrowseComp as open-weights — featured image for AI Insiders
Models / Jun 2

MiniMax M3 beats Opus 4.7 on BrowseComp as open-weights

By Alessandro Benigni
Nvidia's N1X laptop chip targets Intel's last stronghold — featured image for AI Insiders
Models / Jun 2

Nvidia's N1X laptop chip targets Intel's last stronghold

By Alessandro Benigni
Prism ML ships 4B-parameter diffusion that runs on an iPhone — featured image for AI Insiders
Models / Jun 2

Prism ML ships 4B-parameter diffusion that runs on an iPhone

By Alessandro Benigni
Anthropic releases Opus 4.8 with effort controls and cheaper fast mode — featured image for AI Insiders
Models / May 30

Anthropic releases Opus 4.8 with effort controls and cheaper fast mode

By Alessandro Benigni
Microsoft is reportedly building its own AI coding model — featured image for AI Insiders
Models / May 30

Microsoft is reportedly building its own AI coding model

By Alessandro Benigni
MiniMax teases M3 with sparse attention that runs 15.6x faster at long context — featured image for AI Insiders
Models / May 30

MiniMax teases M3 with sparse attention that runs 15.6x faster at long context

By Alessandro Benigni
NVIDIA's gamma-World adds independently controllable multi-agent rollouts — featured image for AI Insiders
Models / May 30

NVIDIA's gamma-World adds independently controllable multi-agent rollouts

By Alessandro Benigni
Sakana Labs claims a block-wise training method that bypasses backprop — featured image for AI Insiders
Models / May 30

Sakana Labs claims a block-wise training method that bypasses backprop

By Alessandro Benigni
Biohub ships open protein AI stack with 6.8 billion sequences — featured image for AI Insiders
Models / May 29

Biohub ships open protein AI stack with 6.8 billion sequences

By Alessandro Benigni
Callstack's Apex specializes a coding model for React Native — featured image for AI Insiders
Models / May 29

Callstack's Apex specializes a coding model for React Native

By Alessandro Benigni
ElevenLabs ships Music v2 with mid-track genre switching — featured image for AI Insiders
Models / May 29

ElevenLabs ships Music v2 with mid-track genre switching

By Alessandro Benigni
NVIDIA's LocateAnything decodes vision-language bounding boxes in parallel — featured image for AI Insiders
Models / May 29

NVIDIA's LocateAnything decodes vision-language bounding boxes in parallel

By Alessandro Benigni
A new repo tracks the shift to native multimodal models — featured image for AI Insiders
Models / May 28

A new repo tracks the shift to native multimodal models

By Alessandro Benigni
Claude Mythos finds OpenAI's Erdos disproof with a cleaner shorter proof — featured image for AI Insiders
Models / May 28

Claude Mythos finds OpenAI's Erdos disproof with a cleaner shorter proof

By Alessandro Benigni
DeepSWE introduces a coding benchmark that separates what SWE-Bench Pro cannot — featured image for AI Insiders
Models / May 28

DeepSWE introduces a coding benchmark that separates what SWE-Bench Pro cannot

By Alessandro Benigni
Microsoft AI's MAI-Image-2.5 enters Arena top 3 for text-to-image — featured image for AI Insiders
Models / May 28

Microsoft AI's MAI-Image-2.5 enters Arena top 3 for text-to-image

By Alessandro Benigni
AlphaProof Nexus solves nine open Erdos problems at a few hundred dollars each — featured image for AI Insiders
Models / May 27

AlphaProof Nexus solves nine open Erdos problems at a few hundred dollars each

By Alessandro Benigni
GPT 5.2 is the only model that can build a hard benchmark — featured image for AI Insiders
Models / May 27

GPT 5.2 is the only model that can build a hard benchmark

By Alessandro Benigni
GPT-5.6 reportedly launches in June with focus on reasoning and agentic flows — featured image for AI Insiders
Models / May 27

GPT-5.6 reportedly launches in June with focus on reasoning and agentic flows

By Alessandro Benigni
On-policy distillation closes the training gap between student and teacher — featured image for AI Insiders
Models / May 27

On-policy distillation closes the training gap between student and teacher

By Alessandro Benigni
Gemini 3.5 Flash (Low) generates 45% fewer tokens than Medium — featured image for AI Insiders
Models / May 26

Gemini 3.5 Flash (Low) generates 45% fewer tokens than Medium

By Alessandro Benigni
Mythos 1 surfaces in cloud security tools ahead of broader release — featured image for AI Insiders
Models / May 26

Mythos 1 surfaces in cloud security tools ahead of broader release

By Alessandro Benigni
Goodfire finds SAE features are map fragments, not atoms of meaning — featured image for AI Insiders
Models / May 22

Goodfire finds SAE features are map fragments, not atoms of meaning

By Alessandro Benigni
Microsoft in talks to supply Maia 200 chips to Anthropic — featured image for AI Insiders
Models / May 22

Microsoft in talks to supply Maia 200 chips to Anthropic

By Alessandro Benigni
Qwen3.7-Max arrives as a closed-weight agent foundation model — featured image for AI Insiders
Models / May 22

Qwen3.7-Max arrives as a closed-weight agent foundation model

By Alessandro Benigni
Alibaba unveils the Zhenwu M890 to chase AI agent workloads — featured image for AI Insiders
Models / May 21

Alibaba unveils the Zhenwu M890 to chase AI agent workloads

By Alessandro Benigni
ByteDance open-sources Lance, a 3B unified model for image and video — featured image for AI Insiders
Models / May 21

ByteDance open-sources Lance, a 3B unified model for image and video

By Alessandro Benigni
LiteFrame cuts video LLM inference latency 35% by shrinking the vision encoder — featured image for AI Insiders
Models / May 21

LiteFrame cuts video LLM inference latency 35% by shrinking the vision encoder

By Alessandro Benigni
Meta AI's WavFlow generates synchronized audio in raw waveform space — featured image for AI Insiders
Models / May 21

Meta AI's WavFlow generates synchronized audio in raw waveform space

By Alessandro Benigni
No-filter pretraining outperforms curated data at scale — featured image for AI Insiders
Models / May 21

No-filter pretraining outperforms curated data at scale

By Alessandro Benigni
OpenAI model disproves Erdos geometry conjecture from 1946 — featured image for AI Insiders
Models / May 21

OpenAI model disproves Erdos geometry conjecture from 1946

By Alessandro Benigni
Stability AI ships Stable Audio 3.0 with open weights and six-minute tracks — featured image for AI Insiders
Models / May 21

Stability AI ships Stable Audio 3.0 with open weights and six-minute tracks

By Alessandro Benigni
AllenAI ships OlmoEarth v1.1 with 3x compute reduction — featured image for AI Insiders
Models / May 20

AllenAI ships OlmoEarth v1.1 with 3x compute reduction

By Alessandro Benigni
Google launches Gemini 3.5 Flash with 4x speed edge over rival frontier models — featured image for AI Insiders
Models / May 20

Google launches Gemini 3.5 Flash with 4x speed edge over rival frontier models

By Alessandro Benigni
NVIDIA open-sources LongLive 2.0 for real-time interactive video — featured image for AI Insiders
Models / May 20

NVIDIA open-sources LongLive 2.0 for real-time interactive video

By Alessandro Benigni
A 1B model that costs $1,500 to train from scratch — featured image for AI Insiders
Models / May 19

A 1B model that costs $1,500 to train from scratch

By Alessandro Benigni
Alibaba's Qwen3.7 enters Arena rankings at positions 13 and 16 — featured image for AI Insiders
Models / May 19

Alibaba's Qwen3.7 enters Arena rankings at positions 13 and 16

By Alessandro Benigni
NVIDIA hand-delivers Vera CPUs to Anthropic, OpenAI, and Oracle — featured image for AI Insiders
Models / May 19

NVIDIA hand-delivers Vera CPUs to Anthropic, OpenAI, and Oracle

By Alessandro Benigni
LLMs Switch Modes Mid-Training in Ways Optimizers Cannot Fix — featured image for AI Insiders
Models / May 19

LLMs Switch Modes Mid-Training in Ways Optimizers Cannot Fix

By Alessandro Benigni
Censorship in Qwen 3.5 Is a Removable Circuit, Not a Knowledge Gap — featured image for AI Insiders
Models / May 19

Censorship in Qwen 3.5 Is a Removable Circuit, Not a Knowledge Gap

By Alessandro Benigni
Featured image for: Anthropic limits Claude Mythos to a closed security consortium
Models / May 19

Anthropic limits Claude Mythos to a closed security consortium

By Alessandro Benigni
▣ The Newsletter

The morning brief for people inside the AI industry.

One email a day, Tuesday through Saturday. We read 400 papers, 60 cap-tables, and every regulator's docket so you don't. The site you're on is the archive, the newsletter is the product.

9,679
Subscribers
31.0%
Open rate
Free
Forever

AI Insiders lives on LinkedIn. Open the newsletter and tap subscribe — new issues land in your LinkedIn feed and inbox.

Subscribe on LinkedIn → Subscribe on Substack →
Free, forever
Unsubscribe in one click
9,679 already in

Subscribe on LinkedIn, or get the same daily brief by email on Substack.