Yapay Zeka Modelleri

Zhipu AI's GLM-5V-Turbo is a breakthrough multimodal AI that transforms design mockups directly into executable front-end code, leveraging Huawei's domestic chip stack to reduce reliance on U.S. technology.

Japan's National Institute of Informatics has released two open-source LLMs—LLM-jp-4 8B and LLM-jp-4 32B-A3B—that claim superior Japanese language performance over OpenAI's gpt-oss-20b. The models are designed to address gaps in regional language understanding.

Tencent has released Omniweaving, a new AI video generation model that integrates a reasoning large language model to improve prompt adherence. The model, built upon HunyuanVideo-1.5, supports a wide range of video creation and editing tasks. This advancement signals a significant step in AI's ability to understand and execute complex creative instructions.

The Technology Innovation Institute's new AI model, Falcon Perception, is poised to revolutionize market segmentation by enabling open-vocabulary grounding and segmentation directly from natural language prompts. This 0.6B-parameter early-fusion transformer moves beyond traditional modular approaches, allowing for more intuitive and dynamic analysis of visual customer data. The breakthrough promises to bridge the gap between complex visual data and strategic marketing insights.

Discrete visual tokens are emerging as a breakthrough in multimodal AI, with companies like美团 (Meituan) pioneering their use to process images and audio as predictive tokens. This approach, combined with new crypto infrastructure, is reshaping how AI understands the real world.

LLM inference costs are projected to decline by over 90% by 2030, driven by breakthroughs in model compression and consensus architectures. New technologies like 1-bit quantization and multi-model voting are accelerating this trend.

Google has released Gemma 4 as a fully open-source AI model under Apache 2.0, enabling powerful local agent workflows with up to 256K context tokens. The release includes four model variants optimized for mobile and GPU environments.

Arcee AI has released Trinity Large Thinking, an open-weight reasoning model under Apache 2.0, enabling transparent long-horizon agent development and tool use. This move challenges proprietary alternatives and reinvigorates U.S.-led open AI innovation.

Google has launched Gemma 4, a new suite of open-weights models with Apache 2.0 licensing, multi-modality, and 140+ language support to compete with Chinese AI models. The license change may be more impactful than performance benchmarks.

Microsoft has launched advanced voice and image models that move beyond traditional large language models, signaling a strategic shift toward multimodal AI systems. The new models promise faster processing and enhanced integration across Microsoft’s ecosystem.

Google DeepMind has launched Gemma 4, a family of open-source vision-capable LLMs with unprecedented intelligence-per-parameter efficiency. These models enable powerful local AI on smartphones and laptops, redefining what's possible with compact, Apache 2.0 licensed systems.

Personalized Group Relative Policy Optimization addresses the limitations of standard RLHF in aligning LLMs with diverse user preferences. Unlike global optimization methods, this new framework accounts for heterogeneous reward distributions across individuals.

Google has released Gemma 4, its most powerful open model family to date, now fully licensed under Apache 2.0. This marks a major shift toward accessible, commercial-friendly AI development.

Open models like MiniMax M2.7 and GLM-5 have crossed a critical threshold, matching closed-source frontier models in agentic performance—file operations, tool use, and complex instruction following—at a fraction of the cost and latency.

Google has unveiled Gemma 4, its most powerful open-source AI model family to date, now freely available under the Apache-2.0 license. The models are optimized for deployment across devices from smartphones to workstations.

Google has officially launched Gemma 4, its latest open-weight AI model, now commercially available under the permissive Apache 2.0 license. The move empowers developers to deploy, modify, and monetize the model without restrictions.

Microsoft has introduced three new foundational AI models designed to transcribe voice, generate audio, and create images—bolstering its position against rivals in the generative AI race. The models, unveiled six months after the formation of its dedicated AI division, mark a strategic pivot toward integrated multimodal intelligence.

Gemma 4 brings advanced multimodal AI to edge devices, while CEEK integrates AI-driven virtual interactions—transforming how users engage with content on-device.

Gemma 4, Google's latest open-source AI model, brings full multimodal capabilities to edge devices—including smartphones and Raspberry Pi—enabling offline, privacy-first AI applications. Developers now have unprecedented control over local deployments.

ACE-Step 1.5 XL is set to launch in the next 48 hours, marking a major leap in text-to-audio generation. The model, built on the ACE-Step framework, promises enhanced multilingual music synthesis and improved fidelity.

Alibaba has launched Qwen3.6-Plus, its third proprietary AI model in days, signaling an aggressive expansion in generative AI. The release follows the debut of Qwen3-Omni and an open-source coding model, reinforcing Alibaba’s dominance in multilingual and multimodal AI.

Alibaba has launched Qwen3.6-Plus, its latest proprietary AI model designed specifically for developers, marking a rapid expansion in its generative AI portfolio. The release underscores Alibaba’s aggressive push to compete in the global AI infrastructure market.

LTX 2.3 has set a new standard in AI video generation with 50fps output at 2688x1664 resolution and zero morphing motion blur. The breakthrough, showcased by a Reddit user, highlights a leap in temporal consistency and visual fidelity.

China's strongest programming model, Qwen3.6-Plus, has been unveiled by Alibaba, delivering performance that closely rivals Claude. The model marks a major leap in agentic AI for enterprise applications.