Release Notes

Follow along with CodingFleet updates.

Nov 19, 2024

Team owners can now delete requests of users in the team.
Teams now have an activity log, logging team creation, invitations, members, and finally request deletions done by owners.
Qwen2.5-Coder 30B from Qwen is added to model options across all tools.
Fixed some minor issues in the chat tool.
Our biggest discount of the year is here! Check the pricing page.

Elite users can now re-generate responses in the chat tool with a different AI model. If you don't like a response, hover over the model dropdown menu of that result, and choose a different model for generation. You can still view all the responses by using the navigation buttons:
We've added Microsoft Authentication alongside Google, GitHub and regular form filling on the signup/login pages.
Claude 3 Haiku is replaced with the newly released Claude 3.5 Haiku from Anthropic.

We've added Teams feature. Everyone can create a team. Team owners or admins can invite members, manage them, and access each member’s usage history and billing details.

Claude 3.5 Sonnet is updated. Now it is even better in coding and many other fields.
We fixed an authentication error in the Ephemeral VM Instances feature.

We've added two more LLMs: Llama 3.2 90b and Mistral Large 2. Elite users can select across all tools, while others can still use them in the code generator tool.
Google updated Gemini 1.5 Pro and Gemini 1.5 Flash models with new endpoints, enhancing their capabilities.

We're introducing the Ephemeral VM instances feature to Elite users, which allows you to spin up a temporary VM and interact with it using either GPT-4o Mini, GPT-4o and Claude 3.5 Sonnet. Particularly useful for code testing, and execution, although there are surely more other use cases. Learn more.
Elite users can now choose from any available LLM in the chat tool.

We've added OpenAI's o1 models, including o1 Preview and o1 Mini. These models "think" before generating a response, making them especially useful for scientific applications. They're available to paid users only: Pro, Unlimited and Elite.
Input character limit is increased for all user tiers: Pro from 32k to 50k, Unlimited from 64k to 100k, and Elite from 128k to 200k.

You can now upload your code files and directories to the code editor in all tools, except the code generator, reducing the burden of manually copying and pasting into our code editor. Drag and drop is also supported.
We've improved the UI for the chat and other pages.

Claude 3.5 Sonnet now outputs 2x more text than before, up to ~20,000 characters (20kB of code) in one generation.
We've added Meta AI's Llama 3.1 models, namely Llama 3.1 70b and Llama 3.1 405b to the model selection dropdown menu.

We've added the URL Fetching feature to GPT-4o, GPT-4o Mini and Claude 3.5 Sonnet. Learn more.
GPT-4o and GPT-4o Mini now can output 4x more text than before, up to ~40,000 characters (40kB of code) in one generation.

We've added the Python Code Interpreter to GPT-4o and GPT-4o Mini for Elite Users. Learn more.
GPT-4o Mini now completely replaces the old GPT-3.5 Turbo.

We've added GPT-4o Mini from OpenAI to the list of AI models.
Users can now share their generations to any one with the link. Choose your desired generation from the usage history, then click on the Share button at the top right of the page, then hit Publish Link and copy it.

We've replaced Claude 3 Sonnet with Claude 3.5 Sonnet and removed Claude 3 Opus, as Claude 3.5 Sonnet now outperforms both in all areas.

Codestral model from MistralAI is added to the list of Large Language Models (LLMs) to be used in most tools.

All users can now choose Google Gemini (Gemini 1.5 Pro and Gemini 1.5 Flash) models among others, in the code generator tool, while Elite users have the option to select from the list of Large Language Models (LLMs) across all tools.
Added the Dark Theme, which you can toggle in the navbar.

Last Updated: nov 2024