Zhemal Khamidun
Head of AI at Alpina Digital, CPO of AlpinaGPT (~30-person engineering & AI team). Founder of Vibe Academy & Hamidun Consulting, author of the AI channel «Готовим ИИшницу» (@jemal_hamidun).
MULTIFINGROUND: LLMs Recognize Accounting Concepts With 99% Accuracy but Miss Facts
A new benchmark on filings from 90 issuers: models identify the concept with up to 99.4% accuracy, yet in 26.6–41.9% of cases they miss at least one related number.
Interlock teaches agents to finish the job: up to +6.64 points on JobBench
The Interlock framework synthesizes long-horizon tasks for LLM agents. Fine-tuning two Qwen3.6 models on their own trajectories gave +4.00–4.50 pp on…
BlenderWorld: GPT-6-Astra Beats Opus-5 in 3D Benchmark with a Score of 79.9
The new BlenderWorld benchmark ran 13 frontier models through 223 3D creation tasks in Blender. GPT-6-Astra leads with 79.9, but spatial coherence still lags.
PrismaCoder: Open 9B and 35B Models Close In on Proprietary Ones in Visual Code Generation
FLOMSS: flow matching splits music into 4 stems and takes 2nd place in listening tests
SliderEdit++ Smooths Out Sliders in AI Image Editors: +18.5% and Up to 20x Faster
JockeyBench: 28 LLM Agent Configurations All Lost Money on Japanese Horse Racing
KANLab: A TensorFlow Playground Fork That Shows KAN Training in the Browser
Lukas Niehaus and colleagues released KANLab: a small Kolmogorov–Arnold network trains in the browser, and every B-spline curve can be viewed and edited by hand.
CyberArena Pits 12 AI Agents Against Each Other in 466 Real-Time Attack-and-Defense Matches
The new CyberArena benchmark makes two LLM agents hack and patch services at the same time. It covers 12 models, 40 tasks and 466 matches, and it breaks the ceiling…
Gemini 4 Argon Found Over Twice as Many Vulnerabilities as Baseline Models in Wiz Tests
Gemini 4 Argon Teaches Apollo 2 Humanoid to Play Cornhole: 7 Hits in a Row
Isomorphic Labs Uses Gemini 4 Argon to Parse 500-Page Patents in a Day Instead of a Month
Habr Author Tests How Many Edits It Takes to Erase an LLM Watermark
IBM Spyre Moves torch.compile From the Homegrown sendnn to Open-Source Inductor
The PyTorch stack for IBM's Spyre accelerator has switched its torch.compile backend from the from-scratch sendnn to backend=inductor from open-source PyTorch.
Meta Brings Muse Wonder to the Enterprise AI Agent Market, but Trust Stands in the Way
Meta unveiled Muse Wonder and entered the enterprise agent segment. The models already complete tasks, but companies aren't ready to trust them with workflows.