Flash

The news, the context, and what changes.

RSS
Flash

Google Open-Sources Native Multimodal Embedding Model EmbeddingGemma 2

Google has released EmbeddingGemma 2, built on the Gemma 4 architecture with 740 million total parameters. The model maps text, code, images, video, and audio into a unified embedding space. Its modular design requires only 270 million parameters for text-only tasks, allowing it to run on the Pixel 11 Pro with approximately 191MB of memory after quantization.

88Z AI editorial1 min read
Flash

Manus Announces Completion of $500 Million Financing

Butterfly Effect, the parent company of Manus, has completed over $500 million in financing, led by Boyu Capital and IDG Capital, with participation from existing shareholders including Tencent. This marks its first funding disclosure since resuming independent operations. Meanwhile, Manus released version 2.0, introducing its self-developed Cascade framework, cloud computers, automated workflows, and the personal intelligent assistant Cue.

88Z AI editorial1 min read
Flash

Anthropic Launches Latest Lightweight Model Claude Haiku 5.5

Anthropic has launched its latest lightweight model, Claude Haiku 5.5, which outperforms DeepSeek V4.1 Flash and GLM-5.3-Flash in benchmarks. Positioned as a competitor to GPT-6 Luna, it offers a lower price point, with a 90% reduction in costs for inputs and outputs within 100,000 tokens. The model is the first to support five levels of effort adjustment, achieving a 72.4% score in the Max setting of the OSWorld 2.1 computer operation test. While powerful, Haiku 5.5 does not replace Sonnet 5.5

88Z AI editorial1 min read
Flash

Mistral AI Open Sources Latest Flagship Large Model Mistral Large 4

Mistral has released its flagship model, Mistral Large 4, featuring 1 trillion parameters and a mixture-of-experts architecture. In third-party evaluations, its intelligence index matches GPT-6 Luna, outperforming GPT-6 and Claude in several rigorous tests, with an 82% cybersecurity vulnerability fix rate. Trained on 3,800 Blackwell chips and supporting over 160 languages, the model's weights will be fully open-sourced by the end of the month.

88Z AI editorial1 min read
Flash

Bilibili Open-Sources Multilingual Translation Model Family Index-Translate

Bilibili's Index LLM team has released the Index-Translate multilingual translation model family, featuring preview weights for 2B, 9B, and 35B-A3B parameter sizes. Built on the Qwen3.5 architecture, the models support translation across 150 languages and have been open-sourced on Hugging Face and ModelScope. Beyond basic instruction following for terminology and format retention, the models are optimized for speech subtitle and dubbing scenarios, allowing translation length adjustment by target

88Z AI editorial1 min read
Flash

Ant Group's Bailin Team Launches New Large Model Ling-3.1-flash

Ant Group's Bailin team has released the new large model Ling-3.1-flash, featuring approximately 560B total parameters and 25B activated parameters, with a 1M context window. Using a hybrid architecture, the model focuses on real-world scenarios such as office work, healthcare, and software development, demonstrating the ability to handle complex tasks like building compilers from scratch.

88Z AI editorial1 min read
Flash

DeepSeek Open-Sources Ascend Infrastructure Components

DeepSeek has officially open-sourced infrastructure components for the Huawei Ascend platform, including TileLang, compute libraries, and distributed communication libraries, corresponding to NVIDIA platform components. TileLang supports DeepSeek V4 series model training, while other components cover core capabilities like matrix operations, with performance nearing hardware limits.

88Z AI editorial1 min read
Flash

LibTV Partners with Douyin AI Creation Wave Plan to Launch 'AI Master Public Course'

LibTV has partnered with the Douyin AI Creation Wave Plan to launch the 'AI Master Public Course'. Featuring Liu Shuheng and Wang Yuchen, creators of the award-winning short film 'He Long', the course deconstructs AI short film creation from narrative, art, and performance perspectives. Focusing on underlying creative logic rather than prompt engineering, it is available on Douyin.

88Z AI editorial1 min read