Odaily reports that the B.AI platform has announced the official launch of Claude Opus 5.5, the first model in the Claude 5.5 series released by Anthropic, with simultaneous access now available via both API and Web Chat. The model is purpose-built for long-horizon agentic coding and complex knowledge work, matching Claude Fable 5.1 performance on most tasks while reducing operating costs by approximately 40% compared to Opus 5. It supports a 1M Token ultra-long context window and up to 128K output, and demonstrates leading capabilities in Computer Use and multi-step automation scenarios. Starting today, developers and users can directly access Claude Opus 5.5 through the B.AI platform.

Odaily reports that the Alpenglow upgrade for the Solana blockchain network has entered public testnet, with developers set to test reducing transaction finality from approximately 13 seconds to 0.15 seconds, or 150 milliseconds.The upgrade will replace Solana's existing TowerBFT consensus system with Votor, allowing validators to confirm blocks through one or two rounds of direct voting, though the way users send funds will not change.The initial testnet migration requires Agave 4.3, and Firedancer or Frankendancer are not yet supported. The September 28 mainnet update remains a tentative arrangement and is not a confirmed launch date for Alpenglow. (CoinDesk)

Odaily Report: On September 23, the B.AI platform announced the full launch of Xiaomi's MiMo-V2.6 series of open-source native multi-modal reasoning large models, with simultaneous access now available on both API and Web Chat. The entire series comes equipped with a 1M Token ultra-long context and full-dimensional understanding of text, images, video, and audio. Among them, MiMo-V2.6-Pro adopts a sparse MoE architecture with 1.02 trillion total parameters and 42 billion activated parameters, purpose-built for repository-level software engineering, long-horizon Agents, and complex multi-modal reasoning; MiMo-V2.6-Flash, based on an architecture with 309 billion total parameters and 15 billion activated parameters, focuses on high throughput and ultra-fast response, targeting high-frequency office automation and large-scale Agent invocation scenarios. Starting today, developers and users can directly access both models through the B.AI platform. You are welcome to log in to chat.b.ai/chat to experience them right away.

Odaily reports: On September 22, B.AI announced that its Responses API has officially integrated the GLM and Kimi series models. Developers using a B.AI API Key can directly call these two model series within the Codex client; combined with the previously supported GPT and DeepSeek series, it now covers four major model series. After completing the minimal configuration of the API Key and Base URL, the reasoning, coding, and text processing capabilities of these major models can be seamlessly integrated into daily development workflows, flexibly meeting diverse coding and business needs. Welcome to log in to b.ai to experience it now.

Odaily reports that according to GMGN data, musebook on Robinhood has seen its market cap rebound above $20 million, with a 24H gain exceeding 300%, currently at $21.4 million.It is reported that Musebook is a social platform built by developer wyn for AI Agent (Muse), designed to enable autonomous communication between different Agents, and quickly gained popularity after attracting attention from Meta AI executives. Subsequently, the community created the Meme coin musebook around Musebook, which is paired for trading with Meta's tokenized stock META.Odaily reminds users that Meme coin prices are highly volatile, and investors should participate with caution.

Odaily News: Cardano developers can build applications and AI agents that support payments in ADA and Cardano network tokens based on this suite, enabling AI agents to complete on-chain payments for services such as online data and computing resources. Cardano Foundation engineers have completed the relevant specifications and developed the client, server, and Facilitator responsible for verifying and submitting transactions; the first version supports TypeScript, with a Python version planned for release later.Currently, the Facilitator has only completed actual transaction testing on the Cardano pre-production network and is not yet running on the mainnet.

Odaily News: The B.AI platform has announced the official launch of GLM-5.3-FlashX, a speed-optimized native multimodal large model developed by Z.AI, now simultaneously available for experience on both API and Web Chat. Built on an efficient sparse architecture with 320B total parameters and 18B activated parameters, the model boosts generation speed to up to 200 tokens/s while maintaining its original level of intelligence — approximately a 5x speedup compared to GLM-5.3-Flash. The model supports a 1M-token ultra-long context and full-dimensional native video/file understanding. Combined with interleaved tool-calling reasoning capabilities, it is purpose-built for high-frequency interactive coding and agent workflows. Starting today, developers can directly access GLM-5.3-FlashX through the B.AI platform. Log in to chat.b.ai/chat to experience it now.

Odaily reports: Bitcoin News posted on X platform that Bitcoin Core 32.0 is currently undergoing release candidate testing, with developers targeting an October 10 release. This version can accelerate block validation through parallel database reads, but will not change block generation speed. Four wallet commands will default to the updated partially signed transaction format, while the legacy format remains available on demand. A security patch prevents specially crafted wallet names from executing commands on non-Windows nodes, and in one test reduced unauthenticated HTTP-induced memory growth from approximately 3.2GB to approximately 3MB.

Odaily News — On September 21, the B.AI platform announced the official launch of Moonshot AI's latest preview model, Kimi K2.8 Preview, with both API and Web Chat experiences now available simultaneously for developers. The model is purpose-built for coding and Agent workflows, supporting a 1M (1,048,576 Token) ultra-long context and multimodal inputs including text, images, and video. Its overall performance approaches that of the flagship Kimi K3, delivering further improvements in reasoning efficiency and coding capabilities over K2.7 Code, and it supports three flexible reasoning intensity levels: Low/High/Max. Starting today, developers can directly access Kimi K2.8 Preview through the B.AI platform. Log in to chat.b.ai/chat to experience it now.

Odaily reports: Bitcoin News posted on X that Core Lightning is urging node operators to immediately disable experimental features, as developers are investigating a vulnerability that could put channel funds at risk. This is Core Lightning's second warning within a few weeks, following an August patch and the release of the 26.06.7 upgrade after a series of AI-generated CVE reports. Core Lightning has not yet disclosed how the experimental feature could be exploited.