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
Launchpad
Be early to the next big token project
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Kuaishou Launches KAT-Coder-Pro V2: Output Doubles to 80K, Focuses on Frontend Aesthetic Generation
According to monitoring by 1M AI News, Kuaishou’s StreamLake platform has released its flagship AI coding model, KAT-Coder-Pro V2, which upgrades in two directions: Agentic Coding and frontend aesthetic generation. In terms of Agentic Coding, V2 is compatible with over 10 mainstream AI coding tools including Claude Code, Cline, Kilo, and OpenCode, and has undergone specialized training and full-chain optimization for the OpenClaw framework, covering scaffold protocol understanding, toolchain invocation, and long-chain execution stability. Frontend aesthetic generation is the key focus of this upgrade. The Kuaishou team has also launched its self-developed “KAT Aesthetic Benchmark,” which employs blind testing by professional designers, setting 10 independent evaluation dimensions specifically for scoring in “no reference image creation” scenarios. Under this benchmark, V2 scored a total of 57.6 in the PPT scenario, leading competitors by 14 to 22 points, with a color matching score of 78; in the Landing Page scenario, it achieved a total score of 59.8, ranking first. Compared to the previous generation baseline, the average score for PPT has increased by 103%, while the Landing Page score has improved by 42%, and individual element scores have risen by 300%. In terms of model specifications, V2 has a context length of 256K, with maximum output increased from 32K in V1 to 80K, supporting streaming output, context caching, MCP, and Function Call. API pricing is set at 2.1 yuan per million tokens for input, 8.4 yuan per million tokens for output, free for cache writing, and 0.42 yuan per million tokens for cache reading. Users can subscribe to use it through the StreamLake platform API or Coding Plan.