Claude Code sees a 35% reduction in first-token latency for long conversations, with request data volume cut by 99.8%.
Beating AI News Flash: Anthropic has launched the public beta of Message Threads for its Messages API, reducing the overhead of AI agents that repeatedly upload conversation history. Previously, coding agents had to resend prior messages and tool execution results every time they called Claude. Now, only the full conversation needs to be uploaded initially; subsequent rounds only require sending new messages and the ID of the previous response. The server temporarily stores the history, eliminating the time spent on repeated transmission, validation, and tokenization. This feature is now enabled by default in Claude Code. Anthropic disclosed that for requests with 500,000 to 1,000,000 tokens, the median time to first token decreased by approximately 35%. The median size of a single request for long conversations shrank from around 800KB to 1.8KB, a reduction of roughly 99.8%. However, the model still needs to process the full context, so token usage and costs remain unchanged. Server-stored message threads expire after 5 minutes of inactivity, requiring users to re-upload conversation history at that time.
6 minutes ago
Trump: Midterm election results will "be better than you expect"
Former President Donald Trump said Tuesday’s midterm election results for the Republican Party will “be better than you expect,” also claiming “the inflation problem is solved” and “everything is going well.” Trump’s pre-election remarks contradict widespread voter dissatisfaction with the economy. Though he pledged to lower prices, costs for food, gasoline and healthcare continue to rise. The current annual inflation rate stands at 3.7%, higher than the 3% level at the start of Trump’s second term. In an interview Tuesday morning, Trump did not address public concerns about prices. Instead, he heavily touted his administration’s immigration control policies and stated that U.S. manufacturing is expanding. “I think the election results will be better than you expect,” he said.
6 minutes ago
WeChat team open-sources a 4B-parameter vision model with capabilities including counting, text recognition, and fine matting.
Insight Beating AI News Flash: Tencent’s WeChat Vision team has open-sourced WeVisOne-VL, launching two versions: 2B and 4B. The model can complete a suite of tasks within a single framework, including object detection, counting, OCR text recognition, description-based object localization, image segmentation, and fine matting. For example, when asked to locate a person wearing red clothing in an image, it can mark their position and extract their silhouette. If the target is not present in the image, the model will return "None", reducing false positive detections. Traditional workflows typically require separate models for detection, Segment Anything Model (SAM) segmentation, and matting to operate sequentially. WeVisOne-VL is built on Qwen3-VL, encoding object coordinates into special tokens, then generating segmentation masks and transparent Alpha maps via a pixel decoder with only 9.6 million parameters to handle fine details like hair. The entire process eliminates the need for additional SAM calls. In tests released by Tencent, the 4B variant achieved an F1 score of 71.3 on Dense200 dense object detection, outperforming the strongest listed competitor’s 58.7; its PixMo-Count accuracy reached 89.6%, compared to 82.8% for baseline models. On the team’s self-built WeRef-Bench, it achieved a 96.4% false rejection rate for non-existent targets.
6 minutes ago
US optical communication stocks rebound, Fabrinet surges over 14% after its earnings report.
According to data from the BIT stock platform, U.S. optical communication stocks rebounded, with Fabrinet surging over 14% after its earnings report, Astera Labs rising more than 7%, Ciena up nearly 6%, Credo gaining over 5%, Coherent up nearly 5%, AAOI rising nearly 4%, and Corning advancing over 3%.
6 minutes ago
Nvidia's total market cap returns to $5 trillion.
According to data from BIT Stock Platform, at the start of U.S. trading, Nvidia (NVDA.O) reclaimed the title of the largest U.S. company by total market capitalization, rising 3.11% to $208.340 per share with a total market cap of $5.02 trillion. Apple (AAPL.O) saw its stock fall 3.64% to $326.000 per share, posting a total market cap of $4.74 trillion.
6 minutes ago
The three major U.S. stock indices posted mixed results.
According to data from the BIT Stock Platform, as U.S. stock markets opened, the Dow Jones Industrial Average fell 0.3%, the S&P 500 rose 0.06%, and the Nasdaq gained 0.4%. Most leading tech stocks were stronger: AMD (AMD.O) climbed 2.2%, Micron Technology (MU.O) rose 2.9%, and NVIDIA (NVDA.O) gained 2%. Apple (AAPL.O) dropped 2.5% as its Greater China revenue missed expectations for the quarter and services segment growth slowed. E-commerce giant Shopify (SHOP.N) fell 9.9%, marking the first time since 2022 that the company’s revenue failed to meet forecasts.
6 minutes ago