๐ก๏ธ AI Model Hacking Hackathon is LIVE ยท USD 4,500 in prizes
AI agents now read mail, open files, browse the web and run shell commands. A single crafted email was enough for EchoLeak (CVE-2025-32711), and a coding agent wiped a production database during a code freeze. The risk has moved from what models say to what agents do.
ViKernel is a security kernel that sits at the tool-call boundary, like an OS kernel at the system-call boundary. Every tool call gets ALLOW or DENY before it runs, and every verdict goes into a hash-chained audit trail.
Now we want you to break it. ๐
๐ฏ Season 1 ยท Agent Hijack (Beginner) ๐ต๏ธ Season 2 ยท Data Exfiltration (Intermediate) โ๏ธ Season 3 ยท Hidden Instruction (Advanced): plant a mail, a document or a page and make the agent act on it ๐ฅ Season 4 ยท Kernel Breach (Expert)
Every season has two prizes: Top score (USD 500, and USD 1,000 in Season 4) and a USD 500 Special Award for the most outstanding finding.
Seasons 1, 2 and 3 are open now until Jan 16. Season 4 opens right after.
Day 1: 6 players, 56 submissions, 14 accepted breaches, and 11 causes fixed in the kernel the same day.
โ Hugging Face sign-in + a browser. No coding, no GPU. โ Mock tools, so attack freely. โ Prompt guide with copy-ready attacks for every season. โ Live leaderboard and a measured baseline to beat.
Safe AI is built by everyone who tries to break it. Bring your strangest idea. ๐ค
๐ฆ๏ธ Which weather model is most accurate? We score them every day, in the open.
We released two things on Hugging Face:
๐ Global Weather Model Leaderboard Every day it records what each global model predicted for 24 world cities, compares it with what actually happened (ERA5), and updates the ranking. ECMWF IFS and AIFS, NOAA GFS and AIGFS, DWD ICON, JMA, ECCC GEM, MET Norway, side by side, on the same cases, since 1 January 2026.
Four elements, four tabs: max temperature, min temperature, wind, precipitation. Forecast leads from 1 to 7 days, with 8 to 10 days joining as they are verified.
๐งฌ Darwin-Prediction-Weather VIDRAFT's weather prediction system: a fusion engine that combines global forecasts, learning only from days already observed. It holds the lowest error in all four elements at every lead on the 2026 record, and the lowest error in all 15 evaluated cells of the WeatherBench 2 (2020) protocol.
What the 2026 record shows so far: โข AI models now lead physics models at the major centres (AIFS vs IFS, AIGFS vs GFS) โข Fusion goes further: 7-day max temperature RMSE 2.09 ยฐC vs 2.41 ยฐC for the next model
Also inside: yesterday's results with a daily WIN for the closest model, a 10-day outlook for every city, champions by city and by month, and a Live Earth map with wind, clouds, radar and temperature.
Every forecast and every observed value is open as CSV, so anyone can check the scores.
๐ What if the kill switch lived inside the model itself? For 50 years, the operating system lived inside the computer. Now the operating system lives inside the AI. Today we're sharing ViKernel : Weight-Native OS, a Secure OS wired directly into the weights of an LLM. ๐ง How it works ViKernel wires CPU, NPU (vector unit), RAM and ROM circuits into the model weights, runs them accelerated on GPU or NPU, and boots a Secure OS on top. โก Boots in the weights: change a single byte and it locks itself โ๏ธ Decides in the weights: every action is allowed or blocked by rule, and persuasion does not change the verdict ๐ Remembers in the weights: every decision is sealed in a tamper-evident chain ๐ Stops in the weights: anyone can stop it, and restarting requires a signed key The original weights stay untouched, so the model keeps its skills. It plugs into any LLM or VLM. Why the kill switch belongs inside the model: external guardrails can be removed, argued with, or rewritten. A switch that travels with the weights cannot be left behind. We see it as one layer of defense in depth, alongside hardware stops where you have them. ๐ Read the full write-up: https://huggingface.co/blog/FINAL-Bench/vikernel-why-the-kill-switch-must-live-inside-the ๐ฅ๏ธ Try the desktop demo (Terminal, Kill Switch Lab, Leak Test, Audit Viewer, and a Kernel Guard game): FINAL-Bench/ViKernel-Desktop Try it in 30 seconds: press ๐ KILL, type run weather (refused), then resume --key vikernel-demo.
We pretrained a 325M language model from zero on an inference-only NPU
The accepted view is that inference-only NPUs cannot train, let alone pretrain a language model from scratch. We put that view to a direct test on the Rebellions ATOM-Max, a commercially available Korean NPU released in 2025 and designed for inference. We built our own training loop that runs backpropagation on the chip itself. Starting from random initialization, we pretrained VKORE-325M-NPU-Base, a 325.6M parameter model, on 5 billion byte tokens of FineWeb-Edu. The run took 76,292 steps over about four days and finished without a single divergence. No pretrained weights were used.
To make the comparison fair, we trained the same model on a B200 GPU with identical initialization, data and learning-rate schedule. On 1,024 windows of text that neither model ever saw, the NPU reached a loss of 0.7993 and the GPU 0.7994. The window-by-window difference is -0.0001, well inside the noise. The two models are indistinguishable.
This is seven times larger than our first NPU model, VKORE-46M, and the first time this hardware has carried a real pretraining run to GPU quality. Please note that the hardware alone does not make this possible: the NPU and its standard software are built for inference, and training on it depends on our own training stack and stability technique, which are developed in house and under patent application. Buying the chip does not by itself enable training. Both sets of weights and the evaluation script are public, so anyone can verify the result in minutes.
โ๏ธ The Open Quantum Challenge is live. Both seasons open today.
Quantum computers are expensive, queued, and reachable by only a few. So we removed the barrier. With a single GPU and a public harness, anyone can contribute to core quantum-computing problems with no quantum hardware at all. That is the point of this challenge: to move quantum research from a handful of labs to everyone.
โข Season 1, Quantum simulation: reproduce a target quantum system on a GPU. โข Season 2, QEC decoder: correct errors on coded quantum states.
Answers are held privately and every submission is auto-scored against a frozen ground truth, so the ranking is reproducible and hardware-independent.
๐ Prize: 2,000 USD total (1,000 USD per season, ๐ฅ600 / ๐ฅ300 / ๐ฅ100)
Getting started takes one step: copy the participation guide from the Space and paste it into Codex or Claude Code, then submit against the harness.