this is a living document, and this post will always show the most recent software update at the top. previous ones can be found below with expandable details.
our 2024 release notes have been archived and can now be found here.
our 2025 release notes have been archived and can now be found here.
6th February 2026 - as fast as a rabbit: intern/creations + dlam now running on claude opus 4.6 
hey everyone - another day, and another great improvement to our agents.
Anthropic just dropped Claude Opus 4.6, their latest state of the art model, so we immediately plugged it straight into rabbit intern/creations and DLAM as part of our agentic stack.
for intern/creations, this means stronger long-context reasoning, better multi-step planning, and more dependable tool use. our initial intern and creations tests have shown improvements in reasoning, planning, design and first-pass output quality.
for DLAM, since Opus 4.6 shows major gains on computer-use benchmarks, you can expect more reliable navigation, fewer mis-clicks, better window handling, and stronger end-to-end task completion as this rolls through DLAM.
overall, this gives our agents more autonomy and stability on real work, whether that’s generating, researching, or directly operating your computer.
here at rabbit, you can expect us to move fast, and keep our products and services in line with the latest advancements in AI. please give both intern/creations and DLAM a look, and let us know your results!
& 
s
team rabbit
5th February 2026 - we just keep clawing away 
hi everyone - thank you so much for all of your support following our epic release last week of DLAM, OpenClaw support on r1, and the tease of Project Cyberdeck.
it’s been an incredible week, and we’re truly awed to see all the love and usage of both DLAM and OpenClaw on r1, as well as all the excitement for Cyberdeck. We’ve done blog posts, press interviews, amazing audio chats on Discord and Twitter Spaces, and the best part? we’re only just getting started 
in keeping up the pace with all these exciting changes, we’re happy to share that an OTA full of improvements is landing now to your r1 devices. This week, we focused on cleaning up some of the rough edges around the OpenClaw integration that we didn’t manage to land in time for the initial release.
here’s what’s new for OpenClaw on r1: 
- we’ve added an option in the r1 settings to reset your OpenClaw connection (settings > device > openclaw reset)
- you can now swipe out of the OpenClaw screen to stop audio playback
- switching to OpenClaw now defaults to showing the home screen
- we’ve fixed the screen flashing whilst waiting for openclaw gateway approval
- if there’s a connection error to your gateway, we’ll stop retrying it after a few attempts
- we’ve removed a false positive on the “secure connection required” message
- we’re allowing user messages to queue up when agent is responding
- you can now swipe up from the OpenClaw home screen to access existing messages
- we’ve renamed where it still said “Moltbot” to “OpenClaw”
in addition to these OpenClaw fixes…
- we’ve fixed the text overlap on OTA release notes (this will apply from the next OTA after this one!)
- we’ve prevented device reset before a countdown timer ends for a factory reset
thanks again, and be sure to keep sharing your feedback, and don’t forget to sign up for early access updates on cyberdeck if you didn’t already!
29th January 2026
Our first update of 2026 is here, and it’s HUGE. So it deserved it’s own thread!
Screenshot done from GMT+1. So it’s coming in less then 17h from now