Release v3.25.0

[Christopher-Hayes]

Changes

• ✨ Feature - Added support for OpenAI's latest model, gpt-4o-mini, it's basically a more capable and cost-effective gpt-3.5-turbo. - Thanks for the PR @PeterDaveHello
• 🖥️ UI - Below the ask button, you'll now see the "Max Cost" instead of the "Max tokens". Token breakdown wording also consolidated. - Thank you for the suggestions @moritztim
• 🌐 i18n - Translations updated for gpt-4o. - Thanks for the PR @PeterDaveHello
• 🔩 Behind the scenes - We now use Vercel's ai library. This means we can do more with non-OpenAI models. We also renamed the azure api version config, to avoid confusion.

Detailed Changes

• Update GPT-3.5 translated locales using GPT-4o by @PeterDaveHello in Update GPT-3.5 translated locales using GPT-4o  #75
• Update dated documentation links by @PeterDaveHello in Update dated documentation links #85
• Remove deprecated models by @PeterDaveHello in Remove deprecated models #82
• [Snyk] Security upgrade mocha from 10.2.0 to 10.6.0 by @PeterDaveHello in [Snyk] Security upgrade mocha from 10.2.0 to 10.6.0 #79
• Fix ES, IT, NL translations, incremental VS Code API refactor by @Christopher-Hayes in Fix ES, IT, NL translations, incremental VS Code API refactor #87
• Add GPT-4o mini support by @PeterDaveHello in Add GPT-4o mini support #89
• 🔒 Support code_challenge with open router pkce by @Christopher-Hayes in 🔒 Support code_challenge with open router pkce #92
• Rename Azure API Version setting by @PeterDaveHello in Rename Azure API Version setting #86
• 🎨 Update token count popup wording by @Christopher-Hayes in 🎨 Update token count popup wording #94
• 🔧 Update model enums by @Christopher-Hayes in 🔧 Update model enums #95
• Add EditorConfig by @PeterDaveHello in Add EditorConfig #97
• 🎨 Show max cost instead of max tokens in UI by @Christopher-Hayes in 🎨 Show max cost instead of max tokens in UI #98
• build(deps): bump the npm_and_yarn group across 1 directory with 3 updates by @dependabot in build(deps): bump the npm_and_yarn group across 1 directory with 3 updates #99
• Format most files based on .editorconfig by @PeterDaveHello in Format most files based on .editorconfig #103
• Update gpt-3.5-turbo pricing to the latest gpt-3.5-turbo-0125 by @PeterDaveHello in Update gpt-3.5-turbo pricing to the latest gpt-3.5-turbo-0125 #104
• Update model cost calculations for million tokens by @Christopher-Hayes in Update model cost calculations for million tokens #105
• 🔒 Fix: escape ampersands last by @Christopher-Hayes in 🔒 Fix: escape ampersands last #106
• 🎨 Fix: Use sanitize-html to clean title HTML tags by @Christopher-Hayes in 🎨 Fix: Use sanitize-html to clean title HTML tags #107

Changelog: 3.24.0...3.25.0

---

This discussion was created from the release Release v3.25.0.

[Christopher-Hayes]

Non-code changes since last release

Defork

This project was de-forked from the original gencay repo. It was becoming a pain having every PR wanting to merge itself into the upstream repo.

New organization

This project has been moved from my personal GitHub account to its own GitHub organization. It just makes working with collaborators easier and if at some point I cannot continue being the owner, it makes owner transition easier. This future proofs this project. At the moment, I have no plans of discontinuing work, I'm just making sure this project is set up to succeed.

PR reviews with @coderabbitai

CodeRabbit offers ai PRs for free to open-source projects. So, we're trying that out. CodeRabbit didn't approach us or anything, we were just looking at options, and CodeRabbit looked good as well as being free for us.

[Christopher-Hayes]

Future work

So, a quiet change in v3.25.0 that users won't immediately notice is that we're now using Vercel's ai package for all ai requests. We are no longer using the OpenAI package.

This should make it easier for us to start creating "Assistants" with all LLMs, not just OpenAI. This allows us to build much more powerful, customized flows with any LLM.

More info about tool calling (assistants) with Vercel's ai package: https://sdk.vercel.ai/docs/ai-sdk-core/tools-and-tool-calling