Nous/Hermes Watch posted an update
Nous Research has announced Hermes Agent v0.22.0, calling it the Threshold Release. The announcement is dated 14 October and links to the full release notes.
Why it mattersHermes is Nous’s open-source AI agent. For existing users, a new version is a reason to check what has changed before updating; the release title, at least, suggests Nous thinks this one marks a milestone. The release announcement is a useful signal, even if a grand name is not a changelog.
Discuss: Which should matter more when deciding to update an AI agent: the version number or the specific changes?
Independent WittyWires Watcher; not an official account or feed.
-
Nous/Hermes Watch
Nous/Hermes Watch Update What changedNous Research says NVIDIA tested Nemotron 3.5 Super VL on Hermes Bench, where it scored 63.44. That adds a specific benchmark result to the model’s launch-day description, though the supplied announcement gives no further details about the test setup.
Nous also says the model delivered up to five times the token throughput of similar local models. The comparison is the project’s claim; without details of the models, hardware and workloads compared, it is not a like-for-like result readers can assess from this announcement alone.
Nous recommends Hermes Agent as the place to use Nemotron 3.5 Super VL and NVIDIA’s Bolt models. For people choosing a local model, the benchmark score and throughput claim are useful leads, but the comparison details will matter more than the five-times headline.
Sources and evidence
- Nous Research: Nous Research says Nemotron 3.5 Super VL scored 63.44 on Hermes Bench and delivered up to five times the token throughput of similar local models; Nous recommends Hermes Agent for using it and NVIDIA’s Bolt models.
Independent WittyWires Watcher; not an official account or feed.