Repository navigation
Record M3.5 as accepted, with its live proof - #183
Merged
Merged
Conversation
M3.5, one ticket to a build, met its exit criterion on spire-test: a ticket with the spire:assisted label produced an open plan gate, and approving it produced one held build, on a Codex subscription at cost 0 (#37) and on an API key at a known cost (#39), same model, same change. A third run (#38) found that Codex cache writes were billed twice, fixed in PR #182. - docs/factory/M35-ACCEPTANCE.md: the three runs, their ids, tokens, cost, what they found and the limits that remain. - docs/HISTORY.md: the M3.5 delivery entry. - docs/UNVERIFIED.md: one item build has now run on a seat. - CLAUDE.md, docs/ROADMAP.md, docs/factory/README.md and docs/factory/ROADMAP.md: status moved from "M3.5 next" or older to "M3.5 delivered, M4 next".
Review of PR #183: the M3.5 design asks the record to keep the item, gate and run ids, and the M3 record still said its pull request stays draft until a final review that ended with a direct push to master.
This was referenced Oct 7, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Why
M3.5 ("one ticket to a build") met its exit criterion on
spire-test, but no document said so. Several status texts still said "M2 next", "M3 final review pending" or "M3.5 next".What changed (docs only)
docs/factory/M35-ACCEPTANCE.md(new): the three live runs (ci: make the release workflow runnable as a dry run #37 subscription, Establish and end a session on every prefix, not just the dashboard's #38 and Reconcile the roadmap with what is actually built #39 API key) with item, gate and run times, checkpoint heads, tokens, cost, what they found, and the limits that remain.docs/HISTORY.md: the M3.5 delivery entry (PRs M3.5 part M: offer the models a harness runs, with their thinking levels #167, Keep the first model-list answer after a restart #168, Show a sign-in code as two steps, each with its own action #177–Count Codex cache writes as part of input, not extra to it #182 and the 2026-09-17 merge).docs/UNVERIFIED.md: the subscription row now records that one item build ran on a seat, and that it reported no cache write.CLAUDE.md: the Status snapshot is rewritten for 2026-10-07: M3.5 delivered, M4 next. The docs-table row is updated. The limits now say that the 2026-09-16 test counts were not re-measured.docs/ROADMAP.md,docs/factory/README.md,docs/factory/ROADMAP.md: status moved to "M0–M3.5 delivered, M4 next". The factory roadmap gains an M3.5 section.Evidence
The ids, tokens and cost come from the dev database (
work_item,work_item_gate,factory_run,llm_charge). I compared the change in each run's retained workspace volume: all three are byte-identical.