Meme coin MARSCOIN’s market cap briefly surged past $240 million, hitting an all-time high.

Beating AI Insight News Brief: OpenAI has added a `configuration_update` feature to GPT-6 Astra. Within the same long conversation via the Responses API, users can now adjust reasoning effort up or down at any time without rewriting the cached prompt prefix. For complex tasks, they can switch to High or Xhigh reasoning levels; for simple tasks, they can drop back to Low, while continuing to reuse the entire preceding long context. Previously, this was a hidden cost trap: reasoning effort was set as an instruction at the start of the context, and switching reasoning mid-conversation would render most cached data unreusable. Codex user tests showed cache hit rates dropped from 98.5% to 65% after switching reasoning levels, only recovering to over 97% once a new cache was rebuilt in the next round. This update is highly practical for long-running AI agents: users can temporarily boost reasoning intensity for complex tasks and lower it back for simple ones, without having to reprocess hundreds of thousands of tokens of prior cache as regular input solely due to the level switch.

According to Onchain Lens monitoring, a single whale completed the following purchases over the past 24 hours: 3.23 million PONS for 913 ETH (valued at approximately $2.26 million); 1.69 million CASHCAT for 160 ETH (approximately $392,000); 7,270 AAVE for 1 million USDC; and 156,450 UNI for 850,000 USDC.

Insight Beating AI Flash News: Artificial Analysis has updated its Intelligence Index to version 4.2, designed to make its rankings more resistant to targeted score manipulation. The weight of private testing has been increased from 20% to 40%, preventing model developers from accessing the full test set in advance and reducing room for benchmark-specific optimization. The new version adds two test modules: AA-Briefcase and GDP.pdf. AA-Briefcase simulates real-world knowledge work projects spanning weeks, requiring models to handle numerous related tasks and thousands of documents. GDP.pdf uses 100 professional PDFs totaling 4,592 pages, challenging models to locate evidence across texts, tables, charts, and footnotes to answer questions. GPQA Diamond has been removed from the overall index. This graduate-level science question benchmark has reached near-saturation: GPT-6 Astra scored the highest at 96.3%, with Gemini 3.8 Flash at 95.3%, making it nearly impossible to distinguish between the top models. After recalculation, Claude Fable 5.1 retains the first place, followed by GPT-6 Astra. Astra notched a score of 55, 4 points higher than GPT-5.6 Sol. Artificial Analysis also noted that Astra uses fewer output tokens compared to other leading models in its tier.

Per TradingBeats monitoring, trader Loracle has shorted CASHCAT and PONS with 3x leverage, holding total short positions worth roughly $20.79 million. Their unrealized loss has since widened to over $6.3 million: the CASHCAT short is valued at ~$5.7 million, with an average entry price of $0.23, incurring an unrealized loss of ~$496,000; the PONS short is worth ~$15.09 million, with an average entry price of $0.54, and an unrealized loss of ~$5.839 million. On-chain Perpetual (Perp) and address analysis tool TradingBeats is now live, supporting real-time Hyperliquid data access, tracing whale operations via address tracking, and delivering comprehensive in-depth analysis.

According to Blockworks data, the Meme coin boom has driven Uniswap to hit all-time highs in daily transaction counts across multiple consecutive trading days. Over the past four days, Uniswap has recorded more than 9 million trades each day, far exceeding its previous levels.

According to GMGN monitoring, crypto KOL Qwerty (@Quanterty) has an unrealized profit of approximately $2.16 million from his Bull Run-related holdings, with a paper return of around 10x. As previously reported, crypto executive He Yi follows Qwerty's account, and the KOL today called for the Bull Run to sweep across the globe.