Pay with

USD

Supports Visa, Mastercard, SEPA & more

Flexible trading, zero fees

Use your crypto for payments worldwide

Basic

Trade crypto freely

Smart access to new on-chain tokens

Magnify your profit with leverage

Convert & Auto-Invest

Trade any size with no fees and no slippage

Get exposure to leveraged positions simply

Pre-Market Trading

Trade new tokens before listing

Advanced

Trade on-chain with Gate Wallet

Smart strategies with automated trading

Follow expert trading strategies

CrossEx Trading

One margin balance, shared across platforms

OTC Bank Transfer

Deposit and withdraw fiat

OTC trades with simple RFQ

Access hundreds of perpetual contracts

One platform for global traditional assets

Event Contracts

Predict price moves and seize opportunities

Trade European-style vanilla options

Unified Account

Maximize your capital efficiency

Introduction to Futures Trading

Learn the basics of futures trading

Join events to earn rewards

Use virtual funds to practice risk-free trading

Stock CFD Derivatives

Access real US stocks and ETFs

Trade quality Hong Kong-listed stocks

Real Korean stocks and top assets

Top Japanese stocks, all in one place

High leverage, 24/7 trading

Stocks Activities

Trade Popular Stocks and Unlock Generous Airdrops

Tokenized Stocks

Backed by real stock assets

Unlock full access to global stock IPOs

Launch

Collect candies to earn airdrops

Quick staking, earn potential new tokens

Hold GT and get massive airdrops for free

Unlock full access to global stock IPOs

Trade on-chain assets and earn airdrops

Earn futures points and claim airdrop rewards

Investment

One-stop digital asset wealth management

Trade & earn at the same time

Earn interest with idle tokens

Stake cryptos to earn in PoS products

Auto-invest on a regular basis

Dual Investment

Profit from market volatility

Earn rewards with flexible staking

Enjoy a Limited-Time 3% Bonus APR

Enjoy a Limited-Time 5% Bonus APR

Limited-time 11% APR on USDT

Top-tier quant strategies

Earn reliable returns from Treasury RWAs

Pledge one crypto to borrow another

One-stop lending hub

Post, share, and explore crypto trends

Live crypto market analysis

Chat with crypto traders

What is happening in crypto

Crypto guides and educational resources

Crypto insights and industry updates

In-depth research and market analysis

More

Promotions

Activity Center

Participate in activities to earn rewards

Invite friends to earn referral rewards

Affiliate Program

Earn exclusive commission rewards

Grow influence and earn airdrops

Real-time platform updates

Huge fee discounts

Asset Management

One‑stop asset management solution

Enterprise digital asset solutions

Developers (API)

Connects to the Gate application ecosystem

Generous API rebate mechanisms

Trade on TradingView

Elevate your trading experience

AI

Your all-in-one conversational AI partner

Use Gate AI directly in your social App

Gate Blue Lobster, ready to go

Gate for AI Agent

AI infrastructure, Gate MCP, Skills, and CLI

Gate Skills Hub

From office tasks to trading, the all-in-one skill hub makes AI even more useful.

Others

Find FAQs and help guides

Grow with the champions

Proof of Reserves

Gate promises 100% proof of reserves

Keep your assets secure

Regulatory Licenses, Registrations, and Approvals

Quasar 438B Exposed as Based on Zhipu GLM-5.2; 174 Token/s Inference Speed on September 3

AI Industry News

2026-09-03 06:17:30

According to Artificial Analysis and CompactifAI changelog, Spain-based Multiverse Computing's Quasar 438B model was revealed to be based on Zhipu's GLM-5.2 on September 3. Artificial Analysis now labels the model as "Quasar 438B (max, based on GLM-5.2)." When Quasar 438B entered beta on August 5, CompactifAI's update notes stated its capabilities were "identical" to GLM-5.2, prompting community questions about whether Quasar is merely a compressed or renamed version. GLM-5.2 has 753B total parameters with 40B active during inference and scores 53 on Artificial Analysis' intelligence index; Quasar reduces this to 438B parameters with a score of 43 but achieves approximately 174 tokens per second inference speed, significantly faster than GLM-5.2's roughly 68 tokens per second.

Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.

Related News

Alibaba Upgrades Qwen3.8-Max to 0902: Programming Tests Rise Across 8 Benchmarks, 3 Exceed Claude Opus 5

Alibaba Updates Qwen3.8-Max, Ranks First in Global Frontend Coding Benchmark at 1691 Points

Abliteration Releases Jailbroken GLM-5.3 in 3 Days, Boosts Exploit Success Rate to 54.4%

Zhipu Reduces Token Inference Costs 80% Since Start of 2026 Using 100K Domestic Chips

China Surpasses US in AI Race With 300M-Parameter Models on Huawei Chips, India Today Reports

In-Depth Analysis

Google Launches Gemini 3.8 Flash After Longest Monthly Losing Streak

Oliver Grant09-03 03:37

Gemini 3.8 Flash Outperforms Claude Opus 5 on Key Benchmarks at Lower Cost

Oliver Grant09-02 18:24

Google Prepares 3.8 Flash AI Model Release as Soon as Wednesday

Oliver Grant09-01 22:33

Comment

0/400

No comments