OpenAI plans to release a new 'open' language model in the coming months | TechCrunchOpenAI plans to release an open-weight language model and is seeking community feedback to improve it.
New AI strategy: OpenAI introduces open-weight modelOpenAI is introducing an open-weight model for developers to refine AI parameters without original training data.
OpenAI lets developers build real-time voice apps - at a substantial premiumOpenAI launched a real-time API for more interactive, spoken-language communication with its LLMs.
OpenAI Releases Reasoning Model o3-mini, Faster and More Accurate Than o1OpenAI o3-mini enhances STEM applications with faster response and improved reasoning capabilities.
OpenAI plans to release a new 'open' language model in the coming months | TechCrunchOpenAI plans to release an open-weight language model and is seeking community feedback to improve it.
New AI strategy: OpenAI introduces open-weight modelOpenAI is introducing an open-weight model for developers to refine AI parameters without original training data.
OpenAI lets developers build real-time voice apps - at a substantial premiumOpenAI launched a real-time API for more interactive, spoken-language communication with its LLMs.
OpenAI Releases Reasoning Model o3-mini, Faster and More Accurate Than o1OpenAI o3-mini enhances STEM applications with faster response and improved reasoning capabilities.
Mistral AI Introduces Saba: Regional Language Model for Arabic and South Indian LanguageMistral AI's Mistral Saba enhances AI performance for Arabic and South Indian languages with specific training datasets.
Google Launches Gemma 3 1B For Mobile and Web AppsGemma 3 1B is a compact language model designed for fast, offline mobile application integration, enhancing user engagement and privacy.
Mistral AI Introduces Saba: Regional Language Model for Arabic and South Indian LanguageMistral AI's Mistral Saba enhances AI performance for Arabic and South Indian languages with specific training datasets.
Google Launches Gemma 3 1B For Mobile and Web AppsGemma 3 1B is a compact language model designed for fast, offline mobile application integration, enhancing user engagement and privacy.
xAI's Grok 3 is available for free to everyone 'for a short time'Grok 3 is now accessible for free, boasting advanced features and capabilities beyond Grok 2.
Meta Open-Sources Large Concept Model, a Language Model That Predicts Entire SentencesMeta's LCM model innovatively uses sentence embeddings to enhance performance and reasoning capabilities beyond traditional token-based approaches.
Send AI to space to spill human secrets to the aliens, scientists urge - but is it worth the risk?AI could be used to communicate with potential alien life by creating a language learning model for extraterrestrials to engage with.
xAI's Grok 3 is available for free to everyone 'for a short time'Grok 3 is now accessible for free, boasting advanced features and capabilities beyond Grok 2.
Meta Open-Sources Large Concept Model, a Language Model That Predicts Entire SentencesMeta's LCM model innovatively uses sentence embeddings to enhance performance and reasoning capabilities beyond traditional token-based approaches.
Send AI to space to spill human secrets to the aliens, scientists urge - but is it worth the risk?AI could be used to communicate with potential alien life by creating a language learning model for extraterrestrials to engage with.
New Tulu 3 claims to beat DeepSeekTülu 3 405B surpasses DeepSeek through advanced Reinforcement Learning techniques.
Meta Releases Llama 3.3: A Multilingual Model with Enhanced Performance and EfficiencyMeta's Llama 3.3 model enhances AI capabilities with improved efficiency, multilingual support, and safety features, setting new benchmarks in reasoning and coding.
New Tulu 3 claims to beat DeepSeekTülu 3 405B surpasses DeepSeek through advanced Reinforcement Learning techniques.
Meta Releases Llama 3.3: A Multilingual Model with Enhanced Performance and EfficiencyMeta's Llama 3.3 model enhances AI capabilities with improved efficiency, multilingual support, and safety features, setting new benchmarks in reasoning and coding.
Introducing LLaVA-Phi: A Compact Vision-Language Assistant Powered By a Small Language Model | HackerNoonLLaVA-Phi showcases the capabilities of smaller language models in multi-modal tasks with only 2.7B parameters.
Meta Releases Llama 3.2 with Vision, Voice, and Open Customizable ModelsLlama 3.2 is Meta's first multimodal language model, allowing interaction with visual and voice data while offering customizable features.
Stanford AI Model Helps Locate Racist Deeds in Santa Clara County | KQEDUsing an open-source language model, the RegLab rapidly scanned millions of deed records, significantly reducing time and cost compared to traditional methods.
Meta Releases Llama 3.2 with Vision, Voice, and Open Customizable ModelsLlama 3.2 is Meta's first multimodal language model, allowing interaction with visual and voice data while offering customizable features.
Stanford AI Model Helps Locate Racist Deeds in Santa Clara County | KQEDUsing an open-source language model, the RegLab rapidly scanned millions of deed records, significantly reducing time and cost compared to traditional methods.
NVIDIA Unveils NVLM 1.0: Open-Source Multimodal LLM with Improved Text and Vision CapabilitiesNVIDIA's NVLM 1.0 is a versatile open-source multimodal language model that enhances performance in both vision-language and text-only tasks after multimodal training.
Mistral AI and NVIDIA Launch Mistral NeMo 12BMistral AI and NVIDIA collaborated on Mistral NeMo 12B; a high-performance, customizable language model ideal for enterprise applications.
NVIDIA Unveils NVLM 1.0: Open-Source Multimodal LLM with Improved Text and Vision CapabilitiesNVIDIA's NVLM 1.0 is a versatile open-source multimodal language model that enhances performance in both vision-language and text-only tasks after multimodal training.
Mistral AI and NVIDIA Launch Mistral NeMo 12BMistral AI and NVIDIA collaborated on Mistral NeMo 12B; a high-performance, customizable language model ideal for enterprise applications.
Training a Bilingual Language Model by Mapping Tokens onto a Shared Character Space | HackerNoonA bilingual Arabic-Hebrew language model using transliteration shows promising effectiveness, outperforming Arabic-only script models despite a smaller training dataset.
Grok-2 Beta Version Released on X PlatformGrok-2 and Grok-2 mini outperform previous models, achieving higher scores in various academic benchmarks and introducing advanced features for users.