chore: bump default LLM models

[eshaffer321]

Summary

• OpenAI default moves gpt-5.4-nano → gpt-5.6-luna (same input price, ~4% cheaper output, newer generation)
• Anthropic default switches from the pinned snapshot claude-haiku-4-5-20251001 to the rolling alias claude-haiku-4-5, so it won't need a manual bump when the snapshot is retired

Test plan

• go build ./...
• go test ./... -race
• ./itemize walmart -dry-run -days 3 -verbose (auth/fetch/matching all fine; no orders had a posted match in this window to exercise categorization directly)
• pre-commit hooks (gofmt, golangci-lint, go vet, go test)

https://claude.ai/code/session_015otsgn2MY6fzR5DkWBkG6n

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 64.95%. Comparing base (f001f1d) to head (e1b94f1).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #92   +/-   ##
=======================================
  Coverage   64.95%   64.95%           
=======================================
  Files          51       51           
  Lines        6979     6979           
=======================================
  Hits         4533     4533           
  Misses       2108     2108           
  Partials      338      338           

Flag	Coverage Δ
unittests	64.95% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
internal/adapters/clients/clients.go	72.72% <ø> (ø)
internal/domain/categorizer/categorizer.go	88.73% <ø> (ø)
internal/infrastructure/config/config.go	81.08% <100.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.