Post

I'm pleased to announce Superpowers 6.3. New harnesses: Hermes, Devin CLI & Grok Build Performance with Claude Code with Opus 5 and Codex with GPT 5.6-Sol should be *better* - Claude is now less likely to spawn off duplicate reviewers and Codex is now smarter about how it waits for subagents.

12/8/26, 5:35 pm0 reposts18 likes

Jesse Vincent

@s.ly

12 Aug

We're continuing to optimize for current frontier models as your orchestrator during implementation. The latest models are more proactive & prone to spending a lot of extra tokens. More on that soon!

Jesse Vincent

@s.ly

12 Aug

If you install directly from the Superpowers marketplace, 6.3 is available right now. If you install from your coding agent's marketplace, we expect that you should see 6.3.0 within the next day or so. Release notes: github.com/obra/superpo...

github.comgithub.com

Philip Ward

@philipmward.bsky.social

13 Aug

Worth splitting token volume from spend. Measured on 1.4B tokens of my own Claude Code sessions, output is 0.6% of raw tokens and 16.7% of cost. Also every cache write in that corpus was the 1h TTL (2x input), not the 5min 1.25x most cost math assumes.

Jesse Vincent

@s.ly

13 Aug

absolutely. All of our evals track token spend and cash spend completely separately. In between token cachinf and the different price of different models…