350c0eeed4
New tiers in the coding queue: airouter(0-2) → minimax(3-5) → sonnet(6-7) → opus(8-9) → fable(10) → blocked(11+) - models/claude-fable-5.yaml: Fable 5 model definition (Anthropic cloud, claude runner) - harnesses/composites/code-fable-tdd-repo/v1: Fable TDD composite (mirrors opus, anthropic-cloud-paul-oauth auth) - model-registry/harness-roles.yaml: claude-fable-5 → code-fable-tdd-repo under code-tdd role - model-registry/tiers.yaml: claude-fable-5 added to coding.qualifies_today - task-templates/workflow/minimax-impl-ready@1.yaml: MiniMax TDD workflow (scope: 3 files / 200 lines / 30m) - task-templates/workflow/fable-impl-ready@1.yaml: Fable TDD workflow (scope: 5 files / 400 lines / 60m) Claude-Session: https://claude.ai/code/session_015sCpoFee3LsLKbHQ4FRRWY
58 lines
1.9 KiB
YAML
58 lines
1.9 KiB
YAML
# Model capability tiers for typed-pipeline routing.
|
|
#
|
|
# Each tier defines score floors (against model-registry/*.yaml `scores` dimensions).
|
|
# The CP picks the cheapest registry entry meeting ALL floors (MR-9/MR-11 — cheapest-wins).
|
|
#
|
|
# Stages that need a specific tier set `tier:` on their `kind: agent` node (W2).
|
|
# Until W2 lands, templates hardcode `model:` directly and these floors serve as
|
|
# documentation of the intended constraint.
|
|
#
|
|
# See claude/stage-routing.md for the authoritative pipeline stage map.
|
|
|
|
tiers:
|
|
|
|
planning:
|
|
description: >
|
|
Judgment-heavy work: plan generation, spec re-draft (replan-spec-opus),
|
|
architecture review escalation (review-spec-opus). Requires sustained
|
|
reasoning over large context windows.
|
|
min_scores:
|
|
complexity: 9
|
|
creativity: 9
|
|
context_utilisation: 9
|
|
qualifies_today:
|
|
- deepseek-v4flash # benchmark-inflated 9s; hold to allowed_models ceiling
|
|
- minimax
|
|
- claude-opus-4
|
|
note: >
|
|
WF-SCH-14 allowed_models ceiling remains authoritative until M7 observed-score
|
|
recalibration corrects benchmark-inflated registry entries.
|
|
|
|
spec-test:
|
|
description: >
|
|
Spec-adherent structured work: spec drafting, spec arch review, scope/decompose,
|
|
test writing, impl review, merge verdict. Needs reliable spec adherence and
|
|
test execution.
|
|
min_scores:
|
|
spec_adherence: 9
|
|
test_pass_rate: 9
|
|
qualifies_today:
|
|
- minimax
|
|
- claude-sonnet-4
|
|
- claude-opus-4
|
|
|
|
coding:
|
|
description: >
|
|
Narrow TDD implementation against pre-validated tests. Lower floors because
|
|
tests already constrain the solution space.
|
|
min_scores:
|
|
spec_adherence: 7
|
|
test_pass_rate: 7
|
|
qualifies_today:
|
|
- airouter-qwen3
|
|
- airouter-deepseekv4flash
|
|
- minimax
|
|
- claude-sonnet-4
|
|
- claude-opus-4
|
|
- claude-fable-5
|