AI posts
AI release notes
Official updates from Anthropic, OpenAI, Google, xAI, Mistral, DeepSeek and Meta, newest first. Open any update for a short AI summary and the full notes.
Compiled by the Toolstash team · Updated
Google · Gemini API
Released gemini-1.5-flash-8b-001, a stable version of our smallest Gemini API model
Released gemini-1.5-flash-8b-001, a stable version of our smallest Gemini API model.
Google · Gemini app
Gemini 1.5 Pro in Gemini Advanced just became more capable
What: Gemini Advanced, with a chat optimized version of 1.5 Pro-002, now provides better and more accurate responses for prompts related to math and exploring complex topics that invite thoughtful conversation. Whether you’re providing detailed, multi-step instructions, or tackling advanced mathematical calculations, Gemini Advanced will help you navigate these challenges with greater ease. Why: We remain committed to rapid development and delivering the best of Gemini to our users. Through continuous model training and valuable user feedback, we’ve further enhanced the capabilities of Gemini 1.5 Pro, continuing to make Gemini Advanced better over time. Upgrade to Gemini Advanced
Google · Gemini app
Priority access with Gemini Advanced: Upload your code files
What: In May, we upgraded Gemini Advanced to a 1 million token context window and added the ability to upload files and get answers and insights about documents. Today, we are adding the ability to upload multiple code files directly from Google Drive or your device to get help understanding, modifying, or debugging your code. We support most code file formats like C, Python, SQL, and more. Why: Elevate your coding experience and accelerate development with Gemini Advanced's 1M token context window, now capable of processing up to 30k lines of code. Get assistance on complex coding questions, saving time and boosting productivity.
Google · Gemini API
Released gemini-1.5-pro-002 and gemini-1.5-flash-002, two new stable versions of Gemini 1.5 Pro and 1.5…
Released gemini-1.5-pro-002 and gemini-1.5-flash-002, two new stable versions of Gemini 1.5 Pro and 1.5 Flash, for general availability. Updated the gemini-1.5-pro-latest model code to use gemini-1.5-pro-002 and the gemini-1.5-flash-latest model code to use gemini-1.5-flash-002. Released gemini-1.5-flash-8b-exp-0924 to replace gemini-1.5-flash-8b-exp-0827. Released the civic integrity safety filter for the Gemini API and AI Studio. Released support for two new parameters for Gemini 1.5 Pro and 1.5 Flash in Python and NodeJS: frequencyPenalty and presencePenalty.
Google · Gemini API
Added thumb-up and thumb-down buttons to model responses, to enable users to provide feedback on the quality…
Added thumb-up and thumb-down buttons to model responses, to enable users to provide feedback on the quality of a response. Added support for Google Cloud credits, which can now be used towards Gemini API usage.
Google · Gemini API
Added an Open in Colab button that exports a prompt – and the code to run it – to a Colab notebook. The…
Added an Open in Colab button that exports a prompt – and the code to run it – to a Colab notebook. The feature doesn't yet support prompting with tools (JSON mode, function calling, or code execution).
Mistral AI · Mistral API
We reduced price on our flagship model, Mistral Large 2
We introduced a free API tier on La Plateforme.
Mistral AI · New model
We released Pixtral and Mistral Small v24.09
We released Pixtral (pixtral-12b-2409) and Mistral Small v24.09 (mistral-small-2409).
Google · Gemini API
Added support for compare mode, which lets you compare responses across models and prompts to find the best…
Added support for compare mode, which lets you compare responses across models and prompts to find the best fit for your use case.
Mistral AI · Mistral API
In le Chat, we added a mitigation against an obfuscated prompt method that could lead to data exfiltration,…
The attack required users to willingfully copy and paste adversarial prompts and provide personal data to the model. No user was impacted and no data was exfiltrated.
Anthropic · Claude API
We’ve added Workspaces to the Developer Console
Workspaces allow you to set custom spend or rate limits, group API keys, track usage by project, and control access with user roles. Read more in our Workspaces blog post.
DeepSeek · New model
deepseek-coder & deepseek-chat Upgraded to DeepSeek V2.5 Model
The DeepSeek V2 Chat and DeepSeek Coder V2 models have been merged and upgraded into the new model, DeepSeek V2.5. For backward compatibility, API users can access the new model through either deepseek-coder or deepseek-chat.
Anthropic · Claude API
We announced the deprecation of the Claude 1 models
Read more in Model deprecations.
Google · Gemini app
Go Live with Gemini, plus 1 more change
What: Starting in English, Gemini Live is a new way to have natural, free flowing conversations with Gemini on your phone. Brainstorm, learn, and practice out loud with real-time spoken responses. Gemini adapts to your conversational style so you can interrupt, ask follow-up questions or come back to the conversation later with ease. Now available in Gemini Advanced, learn more here. Why: Sometimes it’s just easier and faster to talk things out when inspiration or curiosity strikes, whether that's brainstorming a new business idea, simplifying a complex topic, or rehearsing for an important conversation.
Google · Gemini API
Gemini 1.5 Flash supports supplying JSON schema through model configuration
Gemini 1.5 Flash supports supplying JSON schema through model configuration.
Google · Gemini app
Responses in Gemini became faster
What: Gemini responses on our 1.5 Flash model are now up to 50% faster, thanks to major latency improvements we’ve made over the past few weeks. Why: Fast responses enable a more seamless and responsive Gemini experience, helping you get more done across a variety of tasks.
Google · Gemini app
Boost your learning with Gemini's interactive study tools and tailored guidance, plus 1 more change
What: Gemini now makes learning and exploring more engaging and informative with two new capabilities: Learn with more confidence by prompting @OpenStax to pull in trustworthy responses based on Rice University's OpenStax educational resources, complete with citations and links to relevant textbook content. Additionally, test your knowledge on diverse topics with interactive practice questions, featuring feedback and hints along the way. For now, these features are not available to users under 18, outside of the US, or in languages other than English. Why: Partnering with OpenStax, a leading publisher of free, open educational resources, Gemini provides access to a wealth of reliable information. Gemini also transforms learning into an engaging experience with interactive practice quizzes that offer feedback and a variety of topics to cater to diverse learning styles and interests.
Google · Gemini app
Gemini Advanced and Gemini for Google Workspace add-on priority access: Introducing Gems, custom AI experts…
What: Gems are a simple way to customize Gemini towards your specific goals. In a few steps, you can create a Gem by listing specific instructions, details, or even step-by-step workflows. Gems simplify your work, automate repetitive prompts, and empower you to accomplish more with less effort. Your Gems can even act as an expert and adopt your preferred tone or style to help you accomplish your personal goals. Gems are available in most languages for Gemini Advanced users and Gemini for Workspace add-on subscribers. Why: With Gems you can complete tasks more efficiently, in the way you want them done. Whether you need guidance or need to turn your day-to-day productivity and creativity tasks into repeatable workflows, Gems are your custom team of AI experts to help you minimize repetitive prompting and focus on deeper, more creative collaboration with Gemini.
Google · Gemini API
Released the following experimental models: gemini-1.5-pro-exp-0827 gemini-1.5-flash-exp-0827…
Released the following experimental models: gemini-1.5-pro-exp-0827 gemini-1.5-flash-exp-0827 gemini-1.5-flash-8b-exp-0827
Google · Gemini app
Document Upload and Data Analysis now available for Gemini for Google Workspace add-on subscribers, plus 1 more change
What: Users with a qualifying Gemini for Google Workspace add-on subscription can now upload Google Sheets, Google Docs, CSVs, PDFs as well as Excel and Word files from Google Drive or your device. With Document Upload, business users can elevate research and writing tasks, and identify trends within those documents. With Data Analysis, business users can process and explore data for deeper insights as well as create presentation ready charts. Why: Streamline workflows, and process and visualize data across your documents to unlock deeper insights with Gemini
Anthropic · Claude API
We’ve added support for usage of the SDK in browsers by returning CORS headers in the API responses
Set dangerouslyAllowBrowser: true in the SDK instantiation to enable this feature.
Anthropic · Claude API
8,192-token outputs on Claude Sonnet 3.5 are out of beta and no longer require the…
Google · Gemini app
Gemini 1.5 Pro in Gemini Advanced just became more capable
What: Gemini Advanced, with 1.5 Pro, now provides better responses for prompts related to reasoning and coding. Whether you’re looking for Gemini's assistance to tackle multi-step logical challenges that require more expertise, or to analyze and generate more accurate code, Gemini Advanced will help you more efficiently complete tasks. Why: We believe in rapid iteration and bringing the best of Gemini to the world. Through user feedback and model training, we’ve unlocked a more capable version of Gemini 1.5 Pro, and over time Gemini Advanced will continue to get even better. Upgrade to Gemini Advanced
Anthropic · Claude API
Prompt caching is now available as a beta feature in the Claude API
Cache and re-use prompts to reduce latency by up to 80% and costs by up to 90%.
Google · Gemini API
Added support for PDF processing
Added support for PDF processing.
Google · Gemini API
Fine-tuning support released for Gemini 1.5 Flash
Fine-tuning support released for Gemini 1.5 Flash.
DeepSeek · DeepSeek API
API Launches Context Caching on Disk Technology
The DeepSeek API has innovatively adopted hard disk caching, reducing prices by another order of magnitude.
Google · Gemini API
Released gemini-1.5-pro-exp-0801, a new experimental version of Gemini 1.5 Pro
Released gemini-1.5-pro-exp-0801, a new experimental version of Gemini 1.5 Pro.
Google · Gemini app
Gemini is available to teens in more countries and languages
What: Gemini is now available to more teens globally, supporting over 40 languages. To help teenagers navigate Gemini confidently and safely, we’ve put additional policies and safeguards in place, introduced a teen-specific onboarding process, and included an AI literacy guide. Why: Gemini is designed to be a helpful tool for teens. It can spark creativity and encourage them to explore their passions. From understanding school subjects to tackling major milestones like preparing for university, Gemini is here to help teens learn and grow.
Mistral AI · Mistral API
We released version 1.0 of our Python and JS SDKs with major upgrades and syntax changes
Check out our migration guide for details. We released Agents API.