August
- Google introduces Gemini Live; mobile conversational experience for free-flowing dialogues with Gemini
- Introduces 10 new voice options for more natural interactions
- Features mid-conversation interruptions and hands-free operation
- Expanded app integrations including Keep, Tasks, YouTube Music, and upcoming Calendar extension
- Mistral AI introduces La Plateforme, a toolkit for creating custom AI agents, catering to both developers and non-technical users
- Agents API: For deep integration of Mistral’s AI models into applications
- La Plateforme Agent Builder: No-code solution with an intuitive interface for easy AI agent creation
July
- OpenAI begins rolling out Advanced Voice Mode for ChatGPT
- Offers hyper-realistic audio interactions powered by the GPT-4o model
- Features include real-time conversations, ability to interrupt mid-sentence, and detection of emotional intonations
- Google DeepMind's AlphaProof and AlphaGeometry 2 achieve new milestone
- Achieve a silver medal-level score at the 2024 International Mathematical Olympiad
- Solve complex algebra, number theory, and geometry problems
- Demonstrate AI's growing capability in advanced mathematical reasoning
- OpenAI unveils SearchGPT
- AI-powered search engine prototype
- Offers conversational answers with source attribution
- Challenges traditional search paradigms
- Mistral AI unveils Mistral Large 2
- 123B parameter model with 128K token context
- Excels in multilingual tasks, reasoning, and coding
- Offers competitive pricing
- Meta releases Llama 3.1 405B
- Open-source language model with 405 billion parameters
- 128K token context window
- Performance rivaling leading proprietary models
- Mistral AI and NVIDIA release Mistral NeMo
- Compact 12B parameter model with 128K token context
- Designed for efficient local deployment
- Excels in reasoning and coding tasks
- OpenAI replaces GPT-3.5 with GPT-4o mini
- Features 128,000 token context
- 82% performance on MMLU
- Multimodal support and reduced operational costs
- Available for free to ChatGPT users
June
- Anthropic debuts Claude 3.5 Sonnet
- Advanced AI model with superior reasoning, knowledge, and coding skills
- Enhanced speed and new Artifacts feature for improved collaboration
May
- OpenAI releases GPT-4o
- Capable of processing and generating text, images, and audio
- Google releases Gemini 1.5
- Multimodal AI model with a 1 million token context window
April
- Meta releases LLaMA 3
- 70B parameter language model with improved multilingual capabilities
- Udio releases their AI text-to-music platform
- Generative artificial intelligence model designed to produce music from simple text prompts.
- Capable of generating both vocals and instrumentation.
- Opens new avenues for creativity in music production, allowing users to compose complex pieces with ease.
- Mistral AI releases Mixtral 8x22B
- 141B parameter mixture-of-experts model
- Actively uses 39B parameters per inference for enhanced efficiency and performance
March
- Anthropic releases Claude 3
- Outperforms GPT-4 in reasoning and coding benchmarks
February
- Google rebrands Bard as "Gemini"
- Aligns the consumer-facing product name with the underlying model