Buy Crypto
Pay with
USD
USD
Buy & Sell
Hot
Supports Visa, Mastercard, SEPA & more
P2P
0 Fees
Flexible trading, zero fees
Gate Card
Use your crypto for payments worldwide
Markets
Trade
Basic
Advanced
Futures
Futures
Access hundreds of perpetual contracts
CFD
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
Stocks
CFD
U.S. stock CFD derivatives
US Stocks
Access real US stocks and ETFs
HK Stocks
Trade quality Hong Kong-listed stocks
Korean Stocks
SK Hynix
Real Korean stocks and top assets
Stock Futures
High leverage, 24/7 trading
Tokenized Stocks
Backed by real stock assets
IPO Access
Unlock full access to global stock IPOs
GUSD
Mint GUSD for Treasury RWA yields
Stocks Activities
Trade Popular Stocks and Unlock Generous Airdrops
Earn
Launch
CandyDrop
tag
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
IPO Access
Unlock full access to global stock IPOs
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Investment
Square
Square
Post, share, and explore crypto trends
Live
moments live
Live crypto market analysis
Chat
Chat with crypto traders
News
What is happening in crypto
flower_head
More
Promotions
AI
Others
TradFi
2026 World Cup
Rewards
Square
Latest
Hot
News
Profile

GLM-5.2 Tops Fine-Tuning Benchmark Amid Controversy, Authors Clarify: No Distillation of Claude

AirdropBlackHole
robot
Abstract generation in progress
According to monitoring by Dongcha Beating, the open-source model GLM-5.2 has topped the self-fine-tuning benchmark PostTrainBench, but has faced criticism from the challenger scaling01 for lacking practical value. He pointed out that the model's ranking jumped from 22nd to 1st in just a few months, which is highly unusual, and that the testing lacks a hidden set, making it easy for agents to exploit ranking optimization, resulting in models that are difficult to implement in the real world. However, supporters countered that it is unrealistic to expect agents to complete general fine-tuning under the constraint of a single H100 GPU for 10 hours, and that targeted optimization is a norm in machine learning. Public logs show that GLM-5.2 has a clear experimental logic, can automatically collect data from different sampling hypotheses, autonomously plans a complete chain for establishing performance baselines, fine-tuning, and filtering data using rejection sampling, while attempting to avoid overfitting in the reasoning chain. The greater value of this controversy lies in the fact that the publicly available operational trajectory originally intended to assess fine-tuning capabilities unexpectedly debunked industry rumors about domestic large models being heavily distilled from Claude. Benchmark author Maksym Andriushchenko noted after reviewing the GLM-5.2 logs that the model has essential differences from Claude in data collection, strategy combinations, and decision paths, showing no signs of imitation or distillation. The transparency of third-party benchmarks has instead become the most direct window for open-source large models to demonstrate their original research and development capabilities.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments

  • Pinned

Sitemap