Lookonchain APP

App Store

Binance Wallet’s PREDICT presale has raised over 190,000 BNB, with oversubscription standing at approximately 25 times.

1 hours ago

According to relevant pages, the Binance Wallet PREDICT token presale has raised 197,300 BNB, with a target of 7,794.12 BNB, currently oversubscribed by approximately 25 times. Earlier reports noted that the participation threshold for predict.fun’s (PREDICT) Pre-TGE Prime Sale was 60 points, and 15 points will be deducted for joining the current TGE event. The event runs from 16:00 to 18:00 today, with a per-user subscription limit of 6 BNB.

Relevant content

CryptoQuant Founder: After the midterm elections, if inflows into crypto ETFs resume, the market will return to an uptrend.

CryptoQuant founder Ki Young Ju noted in a post that Bitcoin ETF net inflows have been nearly zero this year, yet BTC has only fallen 7%. With ETF demand remaining low and prices stable, large-scale sell-offs by OG whales or miners appear unlikely. Those looking to sell have likely already offloaded their holdings. Ki Young Ju forecasts bullish momentum will return once mid-term uncertainties abate and ETF inflows resume.

7 minutes ago

GitHub Copilot Fusion Officially Launched: Benchmark Scores Close to Claude Code, Testing Costs Cut in Half

Breaking from Beating AI: GitHub officially launched Copilot Fusion, its multi-model scheduling system, at the Universe 2026 conference. The system assigns simple tasks to suitable models; for complex tasks, it first tries cheaper models, switches to more powerful ones if they fail, or enables cross-checking between different models. Tests by independent evaluator Artificial Analysis show Fusion Max scores 60 in overall programming performance, close to Claude Code paired with Opus 5.5 Max’s 62. Fusion Max’s average API cost per task is $5.59, with a runtime of around 36 minutes; by comparison, Claude Code with Opus 5.5 Max costs $11.42 and takes roughly 60 minutes. This means Fusion cuts costs by approximately 51% and reduces runtime by about 40%. While frequent model switching may lead to cache losses, the test recorded a 97% cache hit rate for Fusion, versus 95% for Claude Code. The results are, however, limited to specific benchmark tests and do not guarantee half savings for all tasks in real-world use.

7 minutes ago

100% win rate smart money holds a 25x long ETH position worth $45.34 million.

According to Lookonchain’s monitoring, a smart trader with a 100% win rate and wallet address 0xe813 opened a long position of 19,491 ETH (valued at $45.34 million) using 25x leverage one hour ago. The trader’s unrealized profit has since reached $315,000. To date, this trader has executed 5 trades, all of which were profitable, generating a total profit of $4.6 million.

7 minutes ago

US Central Command: All shipping lanes in the Strait of Hormuz are free of mines.

U.S. Central Command (CENTCOM) stated that in recent days and weeks, Iranian-controlled media have reported Iranian claims that commercial vessels struck mines while transiting the Strait of Hormuz. These claims are false. In fact, all internationally recognized main transit routes in the Strait of Hormuz have been cleared of mines by U.S. forces and will remain mine-free. Thousands of commercial vessels have safely used these routes and continue to do so.

7 minutes ago

Doubao Work opens third-party models to enterprises: Kimi, DeepSeek, and GLM are now accessible via a unified work portal.

Beating AI Insight News Brief: ByteDance’s Doubao Work has announced it is opening third-party large language model (LLM) services to enterprise customers, with the first batch including models such as Moonshot AI’s Kimi K3.1 Pro, Zhipu AI’s GLM-5.3, and DeepSeek’s V4.1 Flash. Enterprise users can now directly switch models in Doubao Work’s chat interface, no longer limited to ByteDance’s in-house Doubao models. Previously, Doubao Work allowed users to connect third-party models on their own, but custom integrations required users to settle fees directly with model providers, while tool calls and output costs on Doubao Work still consumed separate credits. This update adds third-party models directly to the product’s model selection menu, lowering barriers to switching and usage. ByteDance is evolving Doubao Work from an application of its own models into an enterprise AI work hub that supports multiple models. While enterprises’ choice of LLM is one consideration, the core competition lies in which platform they use for daily tasks, as well as permission and cost management.

7 minutes ago

Smart trader 0xe813 opens 25x long on 19,491 $ETH ($45.34M), already up $315K with perfect 5-0 record

Smart trader 0xe813, with a 100% win rate, opened a 25x long on 19,491 $ETH ($45.34M) an hour ago and is now sitting on $315K in unrealized profit. This trader has completed 5 trades, winning all of them and making a total profit of $4.6M.

7 minutes ago

Popular tokens

BitcoinEthereumHyperliquidSolanaTRONBNBTetherAaveXRPPepeFartcoinOndoJupiterUniswapBonkPendleEthenaArbitrumAvalancheLidoChainlinkPolygonDogecoinCardano