🎉 gpt4free 6.7.0 is live!
Below are the emoji-packed release notes generated from the last 34 commits. Feel free to copy-paste them straight into GitHub Releases.
🚀 gpt4free 6.7.0 – “Yuppi-Qwen-mas” Edition
🗓 December 8, 2025
✨ Highlights
- 🖼️ Image caching lands in Yupp, OpenaiChat & Qwen providers – faster re-requests, lighter bandwidth.
- 🎯 LMArena & GeminiCLI get shiny new models:
gemini-3-pro-preview. - 🔐 HuggingChat auth flow completely rewritten – fewer login headaches, clearer errors.
- ⚡ WebSocket media streaming added to OpenaiChat – voice & images stream in real-time.
- 🧠 Smarter parsing of multi-line
<think>&<yapp>blocks in Yupp. - 🌍 UTF-8 enforced everywhere – goodbye Mojibake.
*️⃣ 10 MB stream buffer – because big answers need big pipes.
🏗️ Provider Polish
- Yupp – Account handling, timeout &
ClientResponseErrorrecovery overhauled. - Qwen – Async refactor, token caching, image-upload tweaks.
- OpenaiChat – Stream-session adoption, image-gen
statusparam. - GradientNetwork – Duplicate model entries removed.
- PollinationsAI – Login URL & completion response fixed.
🛠️ Under-the-hood
StreamSessionadopted across Qwen & Yupp for bullet-proof async.- Provider configs updated –
workingflags reset where needed.
🙌 34 commits • 20 files touched • 4 contributors
Huge thanks to @3mora2, @hlohaus & the whole community for keeping the free-LLM dream alive!
💬 Feedback, bugs, memes?
Open an issue or drop by the Discussions tab.
What's Changed
Full Changelog: v6.6.7...v6.7.0
g4f v6.7.0
Download Options
Python Package:
- PyPI:
pip install g4f==v6.7.0
Executables:
- Windows x64:
g4f-windows-v6.7.0-x64.zip - Linux x64:
g4f-linux-v6.7.0-x64 - Linux ARM64:
g4f-linux-v6.7.0-arm64 - macOS x64:
g4f-macos-v6.7.0-x64 - macOS ARM64:
g4f-macos-v6.7.0-arm64
System Packages:
- WinGet:
winget install gpt4free
Docker:
docker pull hlohaus789/g4f:v6.7.0docker pull hlohaus789/g4f:v6.7.0-slim
💻 Having trouble with the .exe from this release?
👉 Check out the Windows launcher for GPT4Free instead:
🔗 https://github.com/gpt4free/g4f.exe 🚀