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
Gate MCP
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
CryptoWorld News reports that DeepSeek has released the V4 series. The flagship model has total parameters of 1.6T, supports a 1M context window, and its inference compute power is only 27% of V3.2.
This series includes two MOE models: v4-pro has total parameters of 1.6T, activating 49B (49 billion) per token; v4-flash has total parameters of 284B (284 billion), activating 13B (13 billion) per token.
Architecture upgrades include a hybrid attention mechanism, which significantly reduces the overhead of long-context processing; the single-token inference FLOPS of v4-pro is only 27% of V3.2.
Pretraining data exceeds 32T tokens, and post-training is conducted in two phases. v4-pro-max claims to be the strongest open-source model currently, with inference performance approaching cutting-edge proprietary models.
The weights are stored using mixed precision of FP4+FP8.