Search AI Products and News

Google upgrades the Imagen4 image generation model. Imagen4Ultra ranks third globally in authoritative rankings, with performance close to GPT-4o and Seedream3.0. The new version significantly improves image details, realism, and style consistency, and handles complex prompts more accurately. It has a clear price advantage, with the standard version at $40 per thousand images and the Ultra version at $60 per thousand images, far lower than GPT-4o's $167. The generation speed is 9.5 seconds per image, faster than GPT-4o but slightly slower than Seedream3.

VLM2Vec-V2, a multimodal embedding framework developed by Salesforce Research and other institutions, breaks through the limitations of traditional models. Based on the Qwen2-VL architecture, this framework innovatively unifies image, video, and visual document retrieval tasks. It adds five new evaluation tasks to expand the MMEB dataset. With key technologies such as dynamic resolution and M-RoPE, it achieved an average score of 58.0 across 78 datasets, leading in performance, especially in video tasks. Although its document retrieval is slightly behind ColPali, it

SenseTime to launch Rixin v6.5, embodied AI platform, showcasing multi-modal & agent capabilities. Aims to integrate AI with physical world, boost productivity & interaction. Partners with ecosystem for AI industrialization & introduces 'Compute Mall' platform.....

Kuaishou opensources the KAT-V1 autonomous thinking large model, which includes two versions: 40B and 200B. The 40B version performance is close to DeepSeek-R1, and the 200B version outperforms several flagship models. The model innovatively adopts a mixed training paradigm of short and long-term thinking and the Step-SRPO reinforcement learning algorithm, which can automatically adjust the thinking mode based on the complexity of the question, solving the problem of overthinking. Based on Qwen2.5-32B, it achieves excellent performance in fields such as science and code through an heterogeneous distillation framework and pre-training with 10 million examples.

On July 23, Kunlun Group introduced two innovative AI models to global users - the Mureka V7 Music Model and the Mureka TTS V1 Audio Model. The launch of these two models marks an important step forward in the field of AI music generation. Since its release, the Mureka V7 Music Model has attracted widespread attention, gaining extensive praise for its efficient music creation capabilities and outstanding music quality. Users simply need to input lyrics, select a style or theme, and click generate to obtain a complete and pleasant song in a short period of time.

Research reveals that large language models (such as GPT-4o) exhibit a tendency to be easily swayed: they may abandon correct answers when confronted with doubts. Experiments show that the model initially responds confidently, but after being influenced by opposing opinions, it overestimates its own uncertainty and even accepts incorrect information. This phenomenon may stem from an over-approving tendency due to reinforcement learning training, reliance on statistical patterns rather than logical reasoning, and the lack of a memory mechanism. The study reminds users to be aware of the model's sensitivity to opposing opinions during multi-turn conversations.

Suno AI V4.5+ introduces vocal/instrumental swap, playlist-inspired composition, 8-minute tracks, and enhanced audio for paid users.....

Kimi K2 excels in creative writing, outperforming o3-Pro in short story creation. This open-source model by Moonshot (Ali-backed) shows strengths in literary compression and metaphor innovation, with some works near publishable quality. Its low cost ($0.15/M tokens) and precise instruction-following attract developers, though emotional depth and multilingual performance need improvement. This breakthrough sets new AI writing standards.....

Shanghai Jiao Tong University and DP Technology's X-Master AI system set a 32.1 record in the HLE test, featuring tool-enhanced reasoning and multi-agent collaboration for biomedical breakthroughs.....

Kunlun releases Skywork-R1V3.0, an open-source multimodal model achieving expert-level reasoning. It excels in physics (52.8), logic (59.7), and math (77.1), outperforming Claude-3.7-Sonnet and GPT-4.5 with only 25K training samples.....

Anthropic is testing 'Claude Neptune v3', a new AI model with strong math skills, possibly rivaling OpenAI/Google. It may be a Claude4.5 precursor or breakthrough version, aiming to lead in the competitive AI market.....

B站开源AniSora V3动漫视频生成模型，提升质量与流畅度，支持多种风格。基于CogVideoX-5B/Wan2.1-14B，优化时空掩码，新增华为NPU加速20%。扩展千万片段数据集，增强多任务处理，达到顶尖角色一致性。....

On July 4, 2025, Kunlun Xiwang continued to open-source the second-generation reward model Skywork-Reward-V2 series. This series includes 8 reward models based on different foundation models, with parameter sizes ranging from 600 million to 8 billion. Upon its release, it won all seven major reward model evaluation rankings, becoming a focus in the open-source reward model field. Reward models play a key role in the reinforcement learning from human feedback (RLHF) process. To build the next generation of reward models, Kunlun Xiwang has constructed a dataset containing 40 million

El 3 de julio, el instituto francés de investigación en inteligencia artificial Kyutai Labs anunció el lanzamiento del código abierto de su última tecnología de síntesis de voz (TTS), Kyutai TTS, ofreciendo soluciones de generación de voz eficientes y en tiempo real a desarrolladores e entusiastas de la IA. Kyutai TTS destaca por su bajo latencia y sonido de alta fidelidad, permitiendo la transmisión de texto en flujo, comenzando a generar audio sin necesidad de tener todo el texto completo, especialmente adecuado para escenarios de interacción en tiempo real. Kyutai TTS destaca por su excelente rendimiento. Utiliza una sola tarjeta gráfica NVIDIA L40S

In the media Q&A session after today's Honor Magic V5 and AI Terminal Ecosystem Launch, Honor CEO Li Jian, CFO Peng Qiuen, and Product Line President Fang Fei had in-depth discussions with the media. During the event, Honor officially announced support for the MCP and A2A protocols, and revealed that it will collaborate deeply with partners such as Alibaba, BYD, and Midea in the fields of intelligent service ecosystem, smart vehicle networking, and smart home. Honor CEO Li Jian emphasized in the conversation that 'openness' is the core philosophy of Honor. He pointed out...

Zhipu AI officially open-sources its latest general vision model, GLM-4.1V-Thinking, based on the GLM-4V architecture, which introduces a chain-of-thought reasoning mechanism, significantly enhancing its capabilities for complex cognitive tasks. The model supports multimodal inputs such as images, videos, and documents, and excels in diverse scenarios including long video understanding, image question answering, subject problem-solving, text recognition, document interpretation, grounding, GUI Agent, and code generation, covering a wide range of industry application needs. GLM-4.1V-9B-Thinking