Anthropic2026-10-07 19:10:00Anthropic launches Claude Haiku 5.5 with roughly 75% lower running costs than its predecessorAnthropic on Oct. 7 introduced Claude Haiku 5.5, calling it the company’s cheapest, fastest and most capable small model so far. The company said average execution costs are about 75% lower than Haiku 4.5, positioning the model for high-volume, cost-sensitive workloads such as summarization, context compression, database queries and classification. Anthropic also said Haiku 5.5 can act as a sub-agent when Opus 5.5 or Sonnet 5.5 handles coding tasks, and described it as the fastest model in its current lineup for latency-sensitive work like real-time customer support and browser use. The release also brings pricing changes and product updates elsewhere in Anthropic’s stack. Haiku 5.5 uses tiered pricing, with much lower rates for prompts under 100,000 tokens, while Sonnet 5.5’s cache read price has been cut in half. Anthropic said Max and Team subscribers will start receiving monthly API credits this week, and it updated its Python and TypeScript SDKs with beta support for computer use and browser use. Haiku 5.5 is now available on Claude Platform as well as AWS, Google Cloud and Microsoft Azure.260
TRAE2026-10-09 10:34:28TRAE merges TraeCode and TraeWork into a unified development platformAI coding tool TRAE has combined its two products, TraeCode and TraeWork, into a single upgraded platform under the TRAE brand. The new platform includes both an Agent mode and an IDE mode, letting users manage projects in one window while assigning multiple AI agents to work in parallel on coding, documentation, testing, and fixes. All outputs are saved automatically. In Agent mode, the platform offers a global workspace designed for multi-agent collaboration on more complex development jobs. The IDE mode keeps conventional code editing and debugging functions, and users can switch between the two through a button in the upper-right corner. TRAE also said the platform includes automated tasks and a template library, while supporting Feishu and WeChat account binding for cross-platform collaboration. The item was carried by Techub and attributed to QbitAI.300
JetBrains2026-10-09 08:06:01JetBrains open-sources Mellum2.1 after posting a higher LiveCodeBench score than Qwen3.5-9BJetBrains has released Mellum2.1, an open-source coding model positioned as an upgrade to Mellum2, which the company introduced in June this year. The new version keeps the same 12 billion total parameters and activates 2.5 billion parameters per generation, but JetBrains said it performs better at identifying code issues, editing files, and checking results. The company said it did not change the model architecture and instead improved performance mainly through reinforcement learning. According to JetBrains, the team ran millions of sandbox tasks across thousands of environments during training, with the model using terminal and file-editing tools inside real code repositories and receiving rewards when tests passed. In JetBrains’ own benchmarks, Mellum2.1 raised its success rate on SWE-bench Verified from 2% to 47%, slightly below Qwen3.5-9B’s 50%. On LiveCodeBench v6, however, Mellum2.1 scored 82%, ahead of Qwen3.5-9B’s 75.4%. JetBrains also said that under heavy-load inference on a single H200, the new model’s output throughput was close to twice that of Qwen3.5-9B. Model weights and a GGUF quantized version are now available on Hugging Face under an Apache 2.0 license for local deployment.310
Monid2026-10-07 03:06:33Monid raises $7.7 million seed round to aggregate about 2,500 APIs for agentsAgent tooling platform Monid has raised $7.7 million in seed funding and is pitching itself as an "OpenRouter for agent tools." The company says developers only need to integrate once, after which an agent can search for, compare, and call tools from different providers on its own. That removes the need to sign up for separate services, manage multiple API keys, or buy individual subscriptions across platforms. Founder Shengkun Ye said Monid now covers about 2,500 APIs. Those APIs span search, SEO, sales leads, e-commerce, stocks, social data, as well as video, image, music, and 3D services, plus agent email and phone tools. According to the description provided, an agent first looks for a suitable tool, then checks the input format and pricing, and finally makes the call directly. Charges are deducted from a unified Monid balance, with different tools billed either by call volume or by returned results.280
OpenAI2026-10-07 02:58:43OpenAI’s 28-day sprint Day 2 brings four updates, with another quota reset still possibleOpenAI product and platform lead Thibault Sottiaux said Day 2 of the company’s 28-day product sprint delivered four updates, while leaving open the possibility of an additional quota reset later in the day. The clearest change for Codex users is that Auto-review is now free for people who sign in with a ChatGPT account and no longer counts against subscription usage. Sottiaux said the feature previously could consume 2% to 10% of a plan’s quota. The other three updates cover OpenAI’s API pricing tiers, a new ChatGPT Meetings plugin, and the public beta launch of the Decisions API. OpenAI has reduced its five paid API usage tiers to three levels: Build, Launch, and Grow. The Meetings plugin can automatically record meetings and organize summaries and action items. The Decisions API is designed to help applications quickly choose a model, a tool, or the next action. Sottiaux also said Day 2 could end up being "four updates, or a quota reset, or both."280
Polygon2026-10-09 05:46:50Polygon Open Money Stack adds support for TRON networkPolygon’s Open Money Stack now supports the TRON network, according to Techub News. The integration is intended to connect assets and liquidity from the TRON ecosystem to Polygon’s underlying infrastructure. Open Money Stack is described as an open financial stack made up of tools and protocols designed to simplify Web3 application development. With the new support in place, developers can build applications using assets on the TRON network. The update points to a broader infrastructure link between the two ecosystems, centered on asset access and developer use cases, as cited in the Techub News report and the post attributed to @justinsuntron.240
Meta2026-10-03 01:29:10Meta open-sources Muse Gadgets and rolls out Home Link for custom AI hardwareMeta has expanded its Muse push beyond software with the launch of Muse Gadgets, an open-source hardware initiative that lets developers build custom peripherals for its personal AI agent. According to the company’s description, developers can run Muse firmware on ESP32 boards or use a Linux SDK to connect more capable devices, including Raspberry Pi 5-class edge hardware. Meta also outlined several reference designs, such as color E Ink displays for schedules and reminders, HDMI display sticks for TVs and monitors, and handheld touch devices that work as portable Muse terminals. At the same time, Meta introduced Muse Home Link, a small USB-C-powered device designed to connect Muse to a home network and communicate with smart home products, including TVs, speakers, and any device that supports HTTPS interfaces. The company said the first production run totals 5,000 units, with shipments expected to begin in the coming weeks. Eligible Muse subscribers will be able to receive the device for free while inventory lasts. Meta framed the move as a way to let developers experiment with new hardware forms around AI agents. Early examples have already appeared, including an E Ink companion display built by Federico Viticci and a wine-cellar tool created with a roughly $50 SenseCAP Watcher device.580
Arkham2026-10-02 11:32:43Arkham updates API docs with role-based usage guidesBlockchain intelligence platform Arkham has updated its API documentation and added step-by-step guides organized by user role and use case, according to Techub News. The new section arranges multiple API calls by task flow and shows requests, responses, and key information points. Arkham said the update is designed to help different types of users complete practical work more efficiently. The update was referenced in a post from @arkham.380
