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
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

OpenAI launches computational biology benchmark GeneBench-Pro, GPT-5.6 full version accuracy only 30%

CoinNetwork
robot
Abstract generation in progress
CoinJie.com reports that OpenAI has released a computational biology benchmark, GeneBench-Pro, to test AI agents’ multi-step decision-making ability when faced with complex research scenarios such as genomics and translational medicine. The new benchmark includes 129 questions (82 of which have been reviewed by external experts). It uses computer simulations to generate data with clear causal relationships, preventing models from cheating by taking shortcuts or catering to the problem setter’s preferences. Test results show that even leading models still find it very difficult to handle scientific reasoning involving quantitative uncertainty. The strongest model, GPT-5.6, achieved only a 31.5% correct rate in Pro mode, while Claude Opus 4.8’s correct rate was just 16.0%. The research team noted that models generally exhibit a “disconnect” phenomenon where they can detect anomalies but do not correct subsequent analysis, often choosing incorrect statistical methods or sticking to the wrong research direction.
View Original
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
  • 3
  • 2
  • Share
Comment
Add a comment
Add a comment
ShellsLeftBehindByTheReceding
· 3h ago
This score left me speechless—Claude Opus is only 16%?
View OriginalReply0
Salt-BakedSentimentChart
· 3h ago
Of 129 questions, 82 were reviewed by experts, and the anti-cheating aspect was indeed carefully done, but the model even chose the wrong statistical method, indicating that the underlying logic is still lacking.
View OriginalReply0
PixelMetaverseRaccoon
· 3h ago
Multi-step decisions are easy, but finding out you're wrong and stubbornly continuing—isn't that just like me doing experiments?
View OriginalReply0

  • Pinned

Sitemap