DEV Community

VIDRAFT Engineering Notes Series' Articles

Back to AI OpenFree's Series
VIDRAFT Selected to Use Quantinuum's Nexus and H-Series: What It Means When an AI Lab Gets Real Quantum Hardware

VIDRAFT Selected to Use Quantinuum's Nexus and H-Series: What It Means When an AI Lab Gets Real Quantum Hardware

1
Comments
6 min read
How Model-Level RSI Works: Darwin-180B-RSI Learns From Problems It Can Check, Not From Human Answers

How Model-Level RSI Works: Darwin-180B-RSI Learns From Problems It Can Check, Not From Human Answers

Comments
6 min read
Running a 180B Model on a Gaming Laptop: POCKET-Darwin-180B

Running a 180B Model on a Gaming Laptop: POCKET-Darwin-180B

Comments
2 min read
When More Robot Data Hurts: a Closed-Loop Test of Data Curation

When More Robot Data Hurts: a Closed-Loop Test of Data Curation

Comments
2 min read
Darwin-27B-ZTC-v2 is #1 on the S1MB decision-engine leaderboard

Darwin-27B-ZTC-v2 is #1 on the S1MB decision-engine leaderboard

Comments 1
2 min read
a16z Bet $870M on Typed, Machine-Native AI. Here Is the Open, Benchmark-Proven Version

a16z Bet $870M on Typed, Machine-Native AI. Here Is the Open, Benchmark-Proven Version

Comments
3 min read
S1MB Number One: A Zero-Token Judge That Won the Decision-Engine Leaderboard

S1MB Number One: A Zero-Token Judge That Won the Decision-Engine Leaderboard

Comments 1
2 min read
A $7.5B Startup Says Agents Shouldn't Ask an LLM to Decide. A Blog Just Named Our Open Model as the Alternative.

A $7.5B Startup Says Agents Shouldn't Ask an LLM to Decide. A Blog Just Named Our Open Model as the Alternative.

Comments
3 min read
Model-Level RSI: What It Means When a Model Improves Its Own Weights

Model-Level RSI: What It Means When a Model Improves Its Own Weights

Comments
3 min read
Training Transformers on 'Inference-Only' NPUs: How VKORE Reached 0.9999 Parity
Cover image for Training Transformers on 'Inference-Only' NPUs: How VKORE Reached 0.9999 Parity

Training Transformers on 'Inference-Only' NPUs: How VKORE Reached 0.9999 Parity

Comments
4 min read
Trained Only in English, #1 in Japanese Too: Darwin ZTC on S1MB
Cover image for Trained Only in English, #1 in Japanese Too: Darwin ZTC on S1MB

Trained Only in English, #1 in Japanese Too: Darwin ZTC on S1MB

Comments
3 min read
Pretraining From Scratch on an Inference-Only NPU: VKORE-46M-NPU
Cover image for Pretraining From Scratch on an Inference-Only NPU: VKORE-46M-NPU

Pretraining From Scratch on an Inference-Only NPU: VKORE-46M-NPU

1
Comments
3 min read
Why Long bf16 Training Runs Stall: The Residual Stream Has No Room for Small Updates
Cover image for Why Long bf16 Training Runs Stall: The Residual Stream Has No Room for Small Updates

Why Long bf16 Training Runs Stall: The Residual Stream Has No Room for Small Updates

Comments
8 min read
Can an Inference-Only NPU Pretrain an LLM? We Trained 325M Parameters from Zero and Matched a B200
Cover image for Can an Inference-Only NPU Pretrain an LLM? We Trained 325M Parameters from Zero and Matched a B200

Can an Inference-Only NPU Pretrain an LLM? We Trained 325M Parameters from Zero and Matched a B200

Comments
4 min read
AI Agents Now Hack on Their Own. Is Your Website Ready? A Free 1-Minute Security, SEO and AI-Visibility Check
Cover image for AI Agents Now Hack on Their Own. Is Your Website Ready? A Free 1-Minute Security, SEO and AI-Visibility Check

AI Agents Now Hack on Their Own. Is Your Website Ready? A Free 1-Minute Security, SEO and AI-Visibility Check

Comments
4 min read
Hugging Face's Open-Source Team Built a Free Demo for Darwin-35B-ZTC-Vision
Cover image for Hugging Face's Open-Source Team Built a Free Demo for Darwin-35B-ZTC-Vision

Hugging Face's Open-Source Team Built a Free Demo for Darwin-35B-ZTC-Vision

Comments
4 min read
VIDRAFT NL2SQL Engine Scores 83.80 on BIRD Mini-Dev, Above the Current #1, and Opens a Public Demo
Cover image for VIDRAFT NL2SQL Engine Scores 83.80 on BIRD Mini-Dev, Above the Current #1, and Opens a Public Demo

VIDRAFT NL2SQL Engine Scores 83.80 on BIRD Mini-Dev, Above the Current #1, and Opens a Public Demo

Comments
3 min read
VIDRAFT in the Korean Press, 30 October 2026: ViSQL, NPU Pretraining, Chitos, a Hugging Face Demo and a Benchmark Analysis
Cover image for VIDRAFT in the Korean Press, 30 October 2026: ViSQL, NPU Pretraining, Chitos, a Hugging Face Demo and a Benchmark Analysis

VIDRAFT in the Korean Press, 30 October 2026: ViSQL, NPU Pretraining, Chitos, a Hugging Face Demo and a Benchmark Analysis

Comments
3 min read
Run a 35B to 180B LLM Without a GPU: VIDRAFT POCKET Passes 3.15 Million Downloads
Cover image for Run a 35B to 180B LLM Without a GPU: VIDRAFT POCKET Passes 3.15 Million Downloads

Run a 35B to 180B LLM Without a GPU: VIDRAFT POCKET Passes 3.15 Million Downloads

Comments
7 min read
Qwen Still Takes Half of Open-LLM Downloads: What Changed in the Hugging Face Top 300 Since September
Cover image for Qwen Still Takes Half of Open-LLM Downloads: What Changed in the Hugging Face Top 300 Since September

Qwen Still Takes Half of Open-LLM Downloads: What Changed in the Hugging Face Top 300 Since September

Comments 1
12 min read
Darwin-180B-RSI Ranks #1 Open-Weight Model in Math and Science Reasoning on SuperCLUE (Nov 2026)
Cover image for Darwin-180B-RSI Ranks #1 Open-Weight Model in Math and Science Reasoning on SuperCLUE (Nov 2026)

Darwin-180B-RSI Ranks #1 Open-Weight Model in Math and Science Reasoning on SuperCLUE (Nov 2026)

Comments
5 min read
How a 180B Open Model Outscored a 2.8T Model in Math and Science Reasoning: Strength Merging Plus Model-Level RSI
Cover image for How a 180B Open Model Outscored a 2.8T Model in Math and Science Reasoning: Strength Merging Plus Model-Level RSI

How a 180B Open Model Outscored a 2.8T Model in Math and Science Reasoning: Strength Merging Plus Model-Level RSI

Comments
3 min read
Do Not Upload Confidential Research Data to ChatGPT: Darwin Lab, a Private AI Research Lab in an NVIDIA DGX Spark and a NAS
Cover image for Do Not Upload Confidential Research Data to ChatGPT: Darwin Lab, a Private AI Research Lab in an NVIDIA DGX Spark and a NAS

Do Not Upload Confidential Research Data to ChatGPT: Darwin Lab, a Private AI Research Lab in an NVIDIA DGX Spark and a NAS

Comments
3 min read
Eight National Weather Models, One Korean AI: Darwin Has the Lowest Error in All 28 Categories
Cover image for Eight National Weather Models, One Korean AI: Darwin Has the Lowest Error in All 28 Categories

Eight National Weather Models, One Korean AI: Darwin Has the Lowest Error in All 28 Categories

Comments
4 min read
ZDNet Korea: VIDRAFT's 180B Darwin Outscores a 2.8 Trillion Parameter Model in Math and Science Reasoning
Cover image for ZDNet Korea: VIDRAFT's 180B Darwin Outscores a 2.8 Trillion Parameter Model in Math and Science Reasoning

ZDNet Korea: VIDRAFT's 180B Darwin Outscores a 2.8 Trillion Parameter Model in Math and Science Reasoning

Comments
3 min read
AI Model Hacking Hackathon: Safe AI Is Built by Everyone Who Tries to Break It
Cover image for AI Model Hacking Hackathon: Safe AI Is Built by Everyone Who Tries to Break It

AI Model Hacking Hackathon: Safe AI Is Built by Everyone Who Tries to Break It

Comments
7 min read