Futures
Access hundreds of perpetual contracts
TradFi
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
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Pre-IPOs
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
Promotions
AI
Gate AI
Your all-in-one conversational AI partner
Gate AI Bot
Use Gate AI directly in your social App
GateClaw
Gate Blue Lobster, ready to go
Gate for AI Agent
AI infrastructure, Gate MCP, Skills, and CLI
Gate Skills Hub
10K+ Skills
From office tasks to trading, the all-in-one skill hub makes AI even more useful.
GateRouter
Smartly choose from 30+ AI models, with 0% extra fees
I saw an interesting benchmark today that says a lot about the capabilities of AI models. PinchBench tested various models on OpenClaw agent tasks, and the results are quite different.
Gemini 3 Flash is at the forefront - with a 95.1% success rate. But what's interesting is that other models are also quite close. minimax-m2.1 achieved 93.6% and kimi-k2.5 achieved 93.4%. Claude Sonnet 4.5 is at 92.7% while GPT-4o is at 85.2%.
The difference between these models doesn't seem very large, but when it comes to agent tasks, these small differences also matter. If you want to know the true efficiency of AI models, looking at such benchmarks is very helpful.