Skip to content

OpenVids 0.5.3

Choose a tag to compare

View all tags
@github-actions github-actions released this 08 Oct 09:17
· 6 commits to main since this release
v0.5.3
3f612fa

OpenVids 0.5.3 — full context windows for the agents.

New

  • Settings › Agents › Context › "Full context windows", on by default. Every model now works with the largest context its provider offers: Claude Haiku 5.5 with 1M tokens instead of 100K, and Codex GPT-5.6 with 1M, GPT-6.1 Sol and GPT-6 Astra with 922K, GPT-6 Sol and Luna with 872K, instead of 272K. Long chats are summarized much later, so the agents keep more of the conversation and the project in view.
  • Some models charge more for very long requests: Claude Haiku 5.5 above 100K tokens, GPT-5.6 above 272K. With the switch on, such a request is billed at the provider's long-context rate. Turn the switch off to keep the previous behaviour: those models then stop at their standard-price limit and no request costs extra.
  • The switch applies from the next turn, in open chats too. It is an OpenVids setting only: your OMP configuration is neither used for this nor changed.
Assets 8
Loading