Great work on the ChartX!
We're interested in understanding its current difficulty. Could you provide evaluation results from GPT-4o?
The paper's results use older models, making it hard to gauge the benchmark's challenge against current SOTA capabilities. Access to other recent models like GPT-4V for community comparison is also now limited.
GPT-4o scores would offer a valuable, up-to-date baseline for the community.
Thanks!
Great work on the ChartX!
We're interested in understanding its current difficulty. Could you provide evaluation results from GPT-4o?
The paper's results use older models, making it hard to gauge the benchmark's challenge against current SOTA capabilities. Access to other recent models like GPT-4V for community comparison is also now limited.
GPT-4o scores would offer a valuable, up-to-date baseline for the community.
Thanks!