News from the AI & ML world
DeeperML - #languagemodels
|
Matthias Bastian@THE DECODER
//
OpenAI has announced the integration of GPT-4.1 and GPT-4.1 mini models into ChatGPT, aimed at enhancing coding and web development capabilities. The GPT-4.1 model, designed as a specialized model excelling at coding tasks and instruction following, is now available to ChatGPT Plus, Pro, and Team users. According to OpenAI, GPT-4.1 is faster and a great alternative to OpenAI o3 & o4-mini for everyday coding needs, providing more help to developers creating applications.
OpenAI is also rolling out GPT-4.1 mini, which will be available to all ChatGPT users, including those on the free tier, replacing the previous GPT-4o mini model. This model serves as the fallback option once GPT-4o usage limits are reached. The release notes confirm that GPT 4.1 mini offers various improvements over GPT-4o mini, including instruction-following, coding, and overall intelligence. This initiative is part of OpenAI's effort to make advanced AI tools more accessible and useful for a broader audience, particularly those engaged in programming and web development. Johannes Heidecke, Head of Systems at OpenAI, has emphasized that the new models build upon the safety measures established for GPT-4o, ensuring parity in safety performance. According to Heidecke, no new safety risks have been introduced, as GPT-4.1 doesn’t introduce new modalities or ways of interacting with the AI, and that it doesn’t surpass o3 in intelligence. The rollout marks another step in OpenAI's increasingly rapid model release cadence, significantly expanding access to specialized capabilities in web development and coding. Recommended read:
References :
Kevin Okemwa@windowscentral.com
//
OpenAI has released GPT-4.1 and GPT-4.1 mini, enhancing coding capabilities within ChatGPT. According to OpenAI on Twitter, GPT-4.1 "excels at coding tasks & instruction following" and serves as a faster alternative to OpenAI o3 & o4-mini for everyday coding needs. GPT-4.1 mini replaces GPT-4o mini as the default for all ChatGPT users, including those on the free tier. The models are available via the “more models” dropdown selection in the top corner of the chat window within ChatGPT.
GPT-4.1 is now accessible to ChatGPT Plus, Pro, and Team users, with Enterprise and Education user access expected in the coming weeks. While initially intended for use only by third-party developers via OpenAI's API, GPT-4.1 was added to ChatGPT following strong user feedback. OpenAI Chief Product Officer Kevin Weil said "We built it for developers, so it's very good at coding and instruction following—give it a try!". These models support the standard context windows for ChatGPT and are optimized for enterprise-grade practicality. GPT-4.1 delivers improvements over GPT-4o on the SWE-bench Verified software engineering benchmark and Scale’s MultiChallenge benchmark. Safety remains a priority, with OpenAI reporting that GPT-4.1 performs at parity with GPT-4o across standard safety evaluations. Recommended read:
References :
@the-decoder.com
//
Google is enhancing its AI capabilities across several platforms. NotebookLM, the AI-powered research tool, is expanding its "Audio Overviews" feature to approximately 75 languages, including less common ones such as Icelandic, Basque, and Latin. This enhancement will enable users worldwide to listen to AI-generated summaries of documents, web pages, and YouTube transcripts, making research more accessible. The audio for each language is generated by AI agents using metaprompting, with the Gemini 2.5 Pro language model as the underlying system, moving towards audio production technology based entirely on Gemini’s multimodality.
These Audio Overviews are designed to distill a mix of documents into a scripted conversation between two synthetic hosts. Users can direct the tone and depth through prompts, and then download an MP3 or keep playback within the notebook. This expansion rebuilds the speech stack and language detection while maintaining a one-click flow. Early testers have reported that multilingual voices make long reading lists easier to digest and provide an alternative channel for blind or low-vision audiences. In addition to NotebookLM enhancements, Google Gemini is receiving AI-assisted image editing capabilities. Users will be able to modify backgrounds, swap objects, and make other adjustments to both AI-generated and personal photos directly within the chat interface. These editing tools are being introduced gradually for users on web and mobile devices, supporting over 45 languages in most countries. To access the new features on your phone, users will need the latest version of the Gemini app. Recommended read:
References :
Nehdiii@Towards AI
//
DeepSeek AI has released its V3-0324 endpoint, offering AI developers access to a powerful 685 billion parameter model. This new endpoint boasts lightning-fast responses and a massive 128K context window, accessible via a simple API key. The model is available without rate limiting at a cost-effective price of $0.88 per 164K output tokens, making it an attractive option for developers seeking high performance at a reasonable price.
Lambda is offering DeepSeek V3-0324 live on its Inference API, providing developers with easy access to this powerful AI model. Towards AI has published a series of articles on DeepSeek V3, including a piece on auxiliary-loss-free load balancing. The DeepSeek V3-0324 highlights includes Major Boost in Reasoning performance, 685B total parameters using a Mixture-of-Experts (MoE) design, Stronger front-end development skills and Smarter tool-use capabilities. However, DeepSeek faces competition from other AI companies, particularly China's Baidu. Baidu recently launched two new AI models, ERNIE X1 and ERNIE 4.5, aiming to compete in the global race for advanced AI. According to TheTechBasic, ERNIE X1 is designed to match DeepSeek R1 in performance but at half the price, while ERNIE 4.5 is capable of handling text, video, images, and audio with improved logic and memory skills. Baidu hopes these new models will help it regain ground against rivals. Recommended read:
References :
Chris McKay@Maginative
//
OpenAI has released its latest AI models, o3 and o4-mini, designed to enhance reasoning and tool use within ChatGPT. These models aim to provide users with smarter and faster AI experiences by leveraging web search, Python programming, visual analysis, and image generation. The models are designed to solve complex problems and perform tasks more efficiently, positioning OpenAI competitively in the rapidly evolving AI landscape. Greg Brockman from OpenAI noted the models "feel incredibly smart" and have the potential to positively impact daily life and solve challenging problems.
The o3 model stands out due to its ability to use tools independently, which enables more practical applications. The model determines when and how to utilize tools such as web search, file analysis, and image generation, thus reducing the need for users to specify tool usage with each query. The o3 model sets new standards for reasoning, particularly in coding, mathematics, and visual perception, and has achieved state-of-the-art performance on several competition benchmarks. The model excels in programming, business, consulting, and creative ideation. Usage limits for these models vary, with o3 at 50 queries per week, and o4-mini at 150 queries per day, and o4-mini-high at 50 queries per day for Plus users, alongside 10 Deep Research queries per month. The o3 model is available to ChatGPT Pro and Team subscribers, while the o4-mini models are used across ChatGPT Plus. OpenAI says o3 is also beneficial in generating and critically evaluating novel hypotheses, especially in biology, mathematics, and engineering contexts. Recommended read:
References :
@www.quantamagazine.org
//
References:
finance.yahoo.com
, Quanta Magazine
,
Researchers are exploring innovative methods to enhance the performance of artificial intelligence language models by minimizing their reliance on direct language processing. This approach involves enabling models to operate more within mathematical or "latent" spaces, reducing the need for constant translation between numerical representations and human language. Studies suggest that processing information directly in these spaces can improve efficiency and reasoning capabilities, as language can sometimes constrain and diminish the information retained by the model. By sidestepping the traditional language-bound processes, AI systems may achieve better results by "thinking" independently of linguistic structures.
Meta has announced plans to resume training its AI models using publicly available content from European users. This move aims to improve the capabilities of Meta's AI systems by leveraging a vast dataset of user-generated information. The decision comes after a period of suspension prompted by concerns regarding data privacy, which were raised by activist groups. Meta is emphasizing that the training will utilize public posts and comments shared by adult users within the European Union, as well as user interactions with Meta AI, such as questions and queries, to enhance model accuracy and overall performance. A new method has been developed to efficiently safeguard sensitive data used in AI model training, reducing the traditional tradeoff between privacy and accuracy. This innovative framework maintains an AI model's performance while preventing attackers from extracting confidential information, such as medical images or financial records. By focusing on the stability of algorithms and utilizing a metric called PAC Privacy, researchers have shown that it's possible to privatize almost any algorithm without needing access to its internal workings, potentially making privacy more accessible and less computationally expensive in real-world applications. Recommended read:
References :
Synced@Synced
//
References:
lambdalabs.com
, MarkTechPost
NVIDIA is pushing the boundaries of language models and AI training through several innovative approaches. One notable advancement is Hymba, a family of small language models developed by NVIDIA research. Hymba uniquely combines transformer attention mechanisms with state space models, resulting in improved efficiency and performance. This hybrid-head architecture allows the models to harness both the high-resolution recall of attention and the efficient context summarization of SSMs, increasing the model’s flexibility.
An NVIDIA research team proposes Hymba, a family of small language models that blend transformer attention with state space models, which outperforms the Llama-3.2-3B model with a 1.32% higher average accuracy, while reducing cache size by 11.67× and increasing throughput by 3.49×. The integration of learnable meta tokens further enhances Hymba's capabilities, enabling it to act as a compressed representation of world knowledge and improving performance across various tasks. These advancements highlight NVIDIA's commitment to addressing the limitations of traditional transformer models while achieving breakthrough performance with smaller, more efficient language models. Lambda is honored to be selected as anNVIDIAPartner Network (NPN) 2025 Americas partner of the year award winner in the category of Healthcare. Artificial intelligence systems designed for physical settings require more than just perceptual abilities—they must also reason about objects, actions, and consequences in dynamic, real-world environments. Researchers from NVIDIA introduced Cosmos-Reason1, a family of vision-language models developed specifically for reasoning about physical environments. NVIDIA, a global leader in AI and accelerated computing, is transforming this field by applyingartificial intelligence (AI)techniques, includinglarge language models(LLMs), to analyze and interpret biological data. Recommended read:
References :
|
BenchmarksBlogsResearch Tools |