fix(config): move built-in model defaults to Opus 5.5 and GPT-6

[zzwong]

Summary

Moves the built-in model tier defaults to the current models and prices Claude Opus 5.5.

Runtime	small	medium	large
Claude CLI	claude-haiku-4-5	claude-sonnet-5	claude-opus-5 → claude-opus-5-5
Codex CLI / OpenAI API	gpt-5.4-mini → gpt-6-luna	gpt-5.4 → gpt-6-sol	gpt-5.5 → gpt-6-sol

• Pricing: adds claude-opus-5-5 standard ($4 in / $20 out, $0.20 cache read, $5 / $8 cache writes) and fast ($8 / $40) rates from Anthropic's pricing page. Opus 5.5 reads cache at 0.05x input instead of 0.1x.
• Fast mode: adds claude-opus-5-5 to the Claude CLI and Anthropic API lists, and gpt-6-astra, gpt-6-sol, gpt-6-luna to the Codex CLI list. Both providers list fast-mode pricing for these models.
• Profiles: anyone with a model_map override is unaffected. Everyone else on these runtimes gets the new models on their next run.

Not in this PR: cost estimates for Codex and OpenAI API runs, which need adapter changes (#621).

Not done yet: a benchmark run comparing the old and new defaults.

Type of Change

• Bug fix
• New feature
• Documentation update
• Repo scaffolding / maintenance

Checklist

• I have read the contributing guide
• My change follows the project structure and conventions
• make tidy passes
• make lint passes
• make test passes
• make build passes

Related Issues

Closes #620

[monit-reviewer]

Automated PR Review

Reviewed commit: 8229e6c09c51
Profile: open-cli-collective-claude-review - Posting as: monit-reviewer

Summary

Reviewer	Findings
go:implementation-tests	0
policies:conventions	0

Reviewer Coverage

• go:implementation-tests — complete (broad); inspected 10 assigned files (11 inspected across reviewers): internal/app/runtime_test.go, internal/cmd/configcmd/configcmd_test.go, internal/cmd/initcmd/initcmd_test.go, internal/config/config.go, internal/config/config_test.go, internal/pipeline/pipeline_test.go, internal/pricing/pricing.go, internal/pricing/pricing_test.go, internal/stagemodel/resolver_test.go, internal/view/config_test.go; skipped: none; constraints: none
• policies:conventions — complete (broad); inspected 6 assigned files (11 inspected across reviewers): README.md, internal/cmd/configcmd/configcmd_test.go, internal/cmd/initcmd/initcmd_test.go, internal/config/config.go, internal/config/config_test.go, internal/view/config_test.go; skipped: none; constraints: none

Inspected files (11)

• README.md
• internal/app/runtime_test.go
• internal/cmd/configcmd/configcmd_test.go
• internal/cmd/initcmd/initcmd_test.go
• internal/config/config.go
• internal/config/config_test.go
• internal/pipeline/pipeline_test.go
• internal/pricing/pricing.go
• internal/pricing/pricing_test.go
• internal/stagemodel/resolver_test.go
• internal/view/config_test.go

0 PR discussion threads considered. 0 summarized; 0 resolved.

---

Completed in 1m 53s | ~$0.93 (est.) | claude-sonnet-5 | cr 0.10.313

Field	Value
Model	claude-sonnet-5
Reviewers	go:implementation-tests, policies:conventions
Engine	claude_cli · claude-sonnet-5
Reviewed by	cr · monit-reviewer
Duration	1m 53s wall · 2m 15s compute
Cost	~$0.93 (est.)
Pricing basis	anthropic-public-2026-09-02
Tokens	42 in / 9.4k out

Per-workstream usage

• orchestrator-selection — claude-sonnet-5
  • In: 6
  • Out: 3.2k
  • Cache read: 29.9k
  • Cache create: 53.2k
  • Cost: ~$0.25 (est.)
  • Duration: 35s
• go:implementation-tests — claude-sonnet-5
  • In: 12
  • Out: 2.5k
  • Cache read: 158.3k
  • Cache create: 69.8k
  • Cost: ~$0.34 (est.)
  • Duration: 34s
• policies:conventions — claude-sonnet-5
  • In: 18
  • Out: 3.2k
  • Cache read: 291.2k
  • Cache create: 51.9k
  • Cost: ~$0.30 (est.)
  • Duration: 53s
• orchestrator-rollup — claude-sonnet-5
  • In: 6
  • Out: 419
  • Cache read: 102.8k
  • Cache create: 4.1k
  • Cost: ~$0.04 (est.)
  • Duration: 12s