Closed h1psta closed 3 months ago
To add to my question, is 4o mini support planned?
Hi @h1psta, the additional tokens are used to create instruction for LLM to interface with Zotero APIs, in addition to answering questions.
I have just created a new release to support GPT-4o and GPT-4o mini. Please give it a try: https://github.com/lifan0127/ai-research-assistant/releases/tag/0.7.1
Thank you!
Ah, I assumed that was the case.. Thanks a lot for the new release, can't wait to try it out with 4o mini :)
Hi @h1psta, the additional tokens are used to create instruction for LLM to interface with Zotero APIs, in addition to answering questions.
I have just created a new release to support GPT-4o and GPT-4o mini. Please give it a try: https://github.com/lifan0127/ai-research-assistant/releases/tag/0.7.1
Thank you!
Thanks a lot @lifan0127, I use a very frequently Aria and I was waiting for this update. Your work is highly valued
@lifan0127 your latest release in unfortunately not compatible with Zotero 7 (which was released last month) . Thank you for your work!
yeah it would be super nice to enable the GPT-4o mini in the pre-release Zotero 7 version :)
@danieltomasz try my quick-fix https://github.com/lifan0127/ai-research-assistant/issues/124
Hi @h1psta Thanks for providing the workaround. I have added a new pre-release for this:
https://github.com/lifan0127/ai-research-assistant/releases/tag/v0.7.1.
@danieltomasz Please manually install this new release.
I have used only about 5 queries like "Summarize paper X" and "Make a coherent narrative connecting papers X, Y and Z" and it already used 20 cents.
Other apps that use OpenAI API keys have a much smaller token footprint, is this expected behavior ?