Models & Labs
OpenAI launches GPT-5.5 Instant for ChatGPT
TechCrunch AIhigh confidence
Why it matters
- →GPT-5.5 Instant reduces hallucinations in critical fields, enhancing reliability.
- →Improved benchmark scores indicate significant advancements in model performance.
- →Enhanced context management offers more personalized and relevant user interactions.
OpenAI has introduced GPT-5.5 Instant, the new default model for ChatGPT, replacing GPT-5.3 Instant. This model aims to reduce hallucinations in sensitive areas such as law and medicine while maintaining low latency. It has shown improved performance on benchmarks like the AIME 2025 math test and the MMMU-Pro multimodal reasoning benchmark. The model also features enhanced context management, allowing it to reference past conversations for personalized responses. Developers can access GPT-5.5 through API, with the previous version available for a limited time.
Read originalMore from TechCrunch AI
© TechCrunch AIMarket & Regulationresearch
AI Leaders Discuss Industry Challenges at Milken Conference
At the Milken Global Conference, key figures from the AI industry highlighted significant challenges facing the sector, including chip shortages and energy constraints. Christophe Fouquet of ASML emphasized the ongoing supply limitations in chip manufacturing, affecting major tech companies like Google and Microsoft. Meanwhile, Google Cloud's Francis deSouza discussed innovative solutions like space-based data centers to address energy issues. Eve Bodnia introduced a novel approach with energy-based models, suggesting a shift from traditional language models. These discussions underscore the evolving landscape and the need for new strategies in AI development.
TechCrunch AI
© TechCrunch AIMarket & Regulationbusiness
xAI Partners with Anthropic for Data Center Deal
xAI's partnership with Anthropic marks a significant shift in its business strategy, transforming from a consumer of compute resources to a provider. By selling the compute capacity of its Colossus 1 data center to Anthropic, xAI immediately monetizes its infrastructure, potentially worth billions. This move suggests a pivot towards becoming a neocloud provider, focusing on data center operations rather than solely on AI model training. While this partnership boosts xAI's financials, it raises questions about the company's long-term ambitions in AI development, especially as it prepares for an IPO alongside SpaceX.
TechCrunch AI
© TechCrunch AIInvestment · $45B
Market & Regulationbusiness
DeepSeek Eyes $45B Valuation in First Funding Round
DeepSeek, a Chinese AI lab, is making a significant impact with its first venture capital round, potentially valuing the company at $45 billion. This rapid increase in valuation reflects the lab's notable advancements in AI, particularly with its large language model that competes with U.S. giants like OpenAI and Anthropic at a fraction of the cost. The funding round, led by China's state investment fund, is a strategic move to develop homegrown AI technologies, especially as DeepSeek's models are optimized for Huawei chips. This initiative not only aims to retain talent but also positions China to enhance its AI capabilities independently of U.S. technology.
TechCrunch AI
More in Models & Labs
Models & Labsmodels
llama.cpp b9041 Release Expands Platform Support
The latest b9041 release of llama.cpp continues its trend of broadening platform compatibility, making it a versatile choice for developers across different environments. Notably, this update includes support for macOS Apple Silicon with KleidiAI enabled, as well as expanded Vulkan and ROCm 7.2 support on Ubuntu. This release doesn't introduce new models but focuses on enhancing the runtime's adaptability across various hardware configurations. By doing so, llama.cpp strengthens its position as a go-to inference runtime for developers seeking flexibility beyond NVIDIA's CUDA ecosystem.
llama.cpp Releases
Models & Labsmodels
Llama.cpp Adds Granite-Speech Support
Llama.cpp's latest update expands its functionality by integrating IBM's Granite-Speech, significantly enhancing its audio processing capabilities. The update features a Conformer encoder with Shaw relative position encoding and a QFormer projector, which efficiently compresses audio data into the LLM embedding space. This ensures precise token-for-token matching with HF transformers on audio clips, demonstrating its robustness. By incorporating these advanced audio processing techniques, llama.cpp becomes a more versatile tool for developers, extending its utility beyond text to include sophisticated audio data handling.
llama.cpp Releases
Models & Labsmodels
llama.cpp b9049 Release Supports MiniCPM-V 4.6
The llama.cpp b9049 release marks a notable step forward by integrating MiniCPM-V 4.6, enhancing the tool's capabilities for developers. This version addresses several bugs and refines features, such as implementing build_attn for flash attention support and improving code style and type checks. The update also extends its reach across various platforms, including macOS, Linux, and Windows, with tailored support for Apple Silicon and Vulkan. These enhancements make llama.cpp a more versatile and reliable tool for developers working with a range of AI models, boosting its performance and usability.
llama.cpp Releases