Happy New Year 2026!
For this first post of the year, a look back at the announcements from late December and early January. On the agenda: OpenAI opens applications for Grove, rolls out “Your Year with ChatGPT”, publishes its 2025 developer recap, and Qwen releases a new version of its image model.
OpenAI Grove: program for founders
January 2, 2026 — OpenAI opens applications for the next cohort of Grove, its support program for early-stage startups.
Applications are open for the next cohort of OpenAI Grove. Grove is a technical program for founders at the earliest stage and includes 5 weeks of hands-on workshops, office hours, and early access.
🇬🇧 Applications are open for the next cohort of OpenAI Grove. Grove is a technical program for founders at the earliest stage and includes 5 weeks of hands-on workshops, office hours, and early access. — @OpenAIDevs on X
What Grove offers
| Aspect | Details |
|---|---|
| Duration | 5 weeks |
| Format | Hands-on workshops |
| Support | Office hours with the OpenAI team |
| Benefit | Early access to new features |
| Target | Founders at the earliest stage |
Your Year with ChatGPT
December 22, 2025 — OpenAI rolls out a personalized recap feature of your year with ChatGPT.
Your Year with ChatGPT! Now rolling out to everyone in the US, UK, Canada, New Zealand, and Australia who have reference saved memory and reference chat history turned on. Just make sure your app is updated.
🇬🇧 Your Year with ChatGPT! Now rolling out to everyone in the US, UK, Canada, New Zealand, and Australia who have reference saved memory and reference chat history turned on. Just make sure your app is updated. — @OpenAI on X
Prerequisites
To access this feature:
- Be in an eligible country (US, UK, Canada, NZ, Australia)
- Have saved memory enabled
- Have chat history enabled
- Update the ChatGPT app
OpenAI for Developers 2025: the recap
January 2, 2026 — The OpenAI Developers team publishes its recap of 2025 advances for developers.
2025: agents got easier to ship. • Reasoning + chat converged in GPT-5.2 • Multimodal by default: PDFs, audio/realtime, images, video • Agent-native stack: Responses API + Agents SDK/AgentKit • Codex: GPT-5.2-Codex + CLI/web/IDE
🇬🇧 2025: agents got easier to ship. Reasoning + chat converged in GPT-5.2, multimodal by default, agent-native stack, and Codex with GPT-5.2-Codex. — @OpenAIDevs on X
2025 Highlights
| Innovation | Description |
|---|---|
| GPT-5.2 | Fusion of reasoning and chat in a single model |
| Multimodal | PDFs, realtime audio, images and video by default |
| Responses API | Native API to build agents |
| Agents SDK/AgentKit | Tools to orchestrate agents |
| GPT-5.2-Codex | Specialized coding model available in CLI, web and IDE |
🔗 OpenAI for Developers in 2025
Qwen-Image-2512: image model update
January 2, 2026 — Qwen announces an update to its image generation model, with immediate support for SGLang and vLLM.
Happy New Year! Huge thanks to the SGLang team for supporting Qwen-Image-2512!
🇬🇧 Happy New Year! Huge thanks to the SGLang team for supporting Qwen-Image-2512! — @Alibaba_Qwen on X
Key Points
- Weight update — Compatible with existing infrastructure
- SGLang — Day-0 support by LMSYS for fast inference
- vLLM — Day-0 support by the vLLM project
- Open-source — Available on Hugging Face
Usage example with SGLang
sglang generate \
--model-path Qwen/Qwen-Image-2512 \
--prompt "Draw a picture of LA in the rain."🔗 Qwen-Image-2512 on Hugging Face
In brief
| Date | Source | News |
|---|---|---|
| Jan 2 | OpenAI | Grove: applications open |
| Jan 2 | OpenAI | 2025 developers recap |
| Dec 22 | OpenAI | Your Year with ChatGPT rolled out |
| Jan 2 | Qwen | Qwen-Image-2512 available |