Skip to content

Model Catalog

Full reference for model selection. See the Agent Roster for the per-agent quick summary.

NeedProviderPool
Primary reasoningOpus 4.6Anthropic MAX (subscription)
Sub-agent codingCodex (GPT-5.3)ChatGPT Plus (unlimited)
Bulk/parallelGPT-4oCopilot Pro (unlimited)
Large doc analysisGemini 3.1 ProGoogle OAuth (free)
Fast lightweightGemini FlashGoogle OAuth (free)
Secondary ClaudeCopilot Sonnet 4.6Copilot Pro (300/mo)
Premium secondaryCopilot Opus 4.6Copilot Pro (100/mo)
TierComplexityExamplesRoute To
T1 — RoutineSingle-file, clear specBug fix, config edit, simple script, CSS tweakCodex CLI → Gemini CLI
T2 — StandardMulti-file, well-definedNew feature (clear spec), module refactorCodex CLI → Gemini CLI → Sonnet
T3 — ComplexArchitectural, ambiguousDesign system work, cross-service coordinationOpus 4.6
T4 — CriticalSecurity, QA, productionSecurity audit, code reviewOpus 4.6

Escalation rule: T1 fails → retry with other free tool → T2. T2 free tools fail → Sonnet. Sonnet fails or task is ambiguous → Opus.

ModelProvider StringCostBest For
Opus 4.6anthropic/claude-opus-4-6Max planT3/T4 coding, orchestration, QA
Sonnet 4.6github-copilot/claude-sonnet-4.6Copilot ProT2 fallback, writing, frontend
Codex (GPT-5.3)openai-codex/gpt-5.3-codexFree (Plus plan)T1/T2 coding, config/infra
Gemini 3.1 Progoogle-gemini-cli/gemini-3.1-pro-previewGoogle OAuthResearch, data processing
Gemini Flashgoogle-antigravity/gemini-3-flashGoogle OAuthHigh-volume triage, budget tasks
GPT-4ogithub-copilot/gpt-4oCopilot ProBulk/parallel
  • Default coding to unlimited/free tools first.
  • Do NOT use OpenRouter for sub-agents — credits are limited.
Terminal window
# Codex CLI — free via ChatGPT Plus. Terminal-Bench 77.3%. Best for T1/T2.
codex --model gpt-5.3-codex --approval-mode full-auto --quiet "task"
# Gemini CLI — free. 1M+ context. Good for T1/T2 + research.
gemini -m gemini-3.1-pro -ni --sandbox permissive "task"
# Claude Code CLI — Max plan. Use for T2/T3.
claude -p --dangerously-skip-permissions --model sonnet --max-budget-usd 3 "task. Build. Fix errors. Commit."