feat: claude-opus-5 joins the model catalog as the reasoning default - #52
Merged
Merged
Conversation
Anthropic shipped Claude Opus 5 (claude-opus-5, $5/$25 per Mtok — same price as Opus 4.8 with ~2x Frontier-Bench performance). Same price + strictly better means it supersedes opus-4-8 rather than sitting beside it: every default route and escalation rung that pointed at claude-opus-4-8 (architecture, design, security-analysis, debugging, and the implementation/testing ladders) now points at claude-opus-5, and the suggested aqe fallback chain follows. Tier placement: a tier above Opus 4.8, NOT Mythos-class — claude-fable-5 stays the flagship/escalation ceiling (Opus 5 lands within ~0.5% of Fable on coding/agentic benchmarks at about half the cost per task, but trails Mythos-class models on frontier domains). claude-opus-4-8 stays in the catalog under a new 'prior' tier for pinned configs. No upstream changes needed: aqe's model-mapping normalization (ADR-043) passes unknown model IDs through verbatim, and ruflo's providers configure -m takes a free-form model string. The escalatePolicy fixture that pinned the old rung literal now derives it from DEFAULT_ROUTES so it cannot drift on future model bumps.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Anthropic shipped Claude Opus 5 — API ID
claude-opus-5, $5/$25 per Mtok (identical to Opus 4.8) with ~2× Frontier-Bench performance. Same price + strictly better means it supersedesclaude-opus-4-8rather than sitting beside it.MODEL_CATALOG:claude-opus-5added as the recommendedreasoningpick;claude-opus-4-8retained under a newpriortier for pinned configs; catalog re-verified 2026-07-24.DEFAULT_ROUTES: architecture, design, security-analysis, debugging, and the implementation/testing cross-vendor escalation rungs now route toclaude-opus-5.AQE_FALLBACK_CODEX_SUGGESTION), CLI help/prompts, sentinel-block templates (claude/providers-reference.md,claude/aqe-reference.md— heal into~/.claude/CLAUDE.mdon nextak sync), anddocs/PROVIDERS.mdall follow.claude-fable-5remains the flagship/escalation ceiling (Opus 5 is within ~0.5% of Fable on coding/agentic benchmarks at ~half the cost per task, but trails Mythos-class models on frontier domains).Why no upstream changes
Grounded in rUv's source: agentic-qe's model-mapping normalization (ADR-043,
src/shared/llm/model-mapping.ts) passes unknown model IDs through verbatim, andruflo providers configure -maccepts a free-form model string — so the new ID needs curation only, no plumbing.Tests
escalatePolicyfixture that pinned the old rung literal now derives it fromDEFAULT_ROUTES.implementation.escalate[0]so it cannot drift on future model bumps.🤖 Generated with Claude Code