Model Selection

Hosted providers (Gemini, Codex) auto-select a sensible default model with automatic fallback to a lighter model on quota errors. Ollama runs locally and uses exactly the model you pull — it never substitutes another. Most users should never override the model parameter — the defaults are tuned for quality.

Defaults & Fallbacks

Provider	Default	Fallback	Trigger
Gemini	gemini-3.1-pro-preview	gemini-3.5-flash	RESOURCE_EXHAUSTED quota error or "exhausted your capacity" pattern (ADR-044)
Codex	gpt-5.5	gpt-5.5-mini	Quota errors (rate_limit_exceeded, 429, insufficient_quota) (ADR-028, bumped per ADR-067)
Ollama	qwen3.6:27b	— (none)	Local — no fallback; a missing model returns a clear ollama pull error

For Gemini and Codex, the fallback fires automatically inside the executor — your client sees a successful response with usage.fellBack: true in the structured output, and a [Gemini stats: ... fell back] annotation in the formatted text. Ollama never falls back, so its fellBack is always false.

Choosing a Provider

Different providers excel at different things. Pick by what you're doing, not by which is "best":

Task	Suggested provider	Why
Targeted code reasoning, refactor critique	Codex	GPT-5.5's strength is dense code reasoning at moderate context size
Private / air-gapped analysis	Ollama	Runs locally, nothing leaves your machine
Subscription-backed second opinion, larger context	Antigravity	agy via your Google AI Pro/Ultra plan — the Gemini CLI successor
Whole-codebase review (enterprise seats)	Gemini	1M+ token context fits what others can't — enterprise-gated from 2026-06-18
"What do they all think?" comparison	Multi-LLM (multi-llm tool or /compare skill)	Parallel dispatch, see all responses side-by-side
Code review with verified findings	/multi-review skill	Antigravity + Codex in parallel, then verifies each finding against source

Overriding the Model

Pass model explicitly when you have a reason to:

Use ask-llm with provider gemini and model gemini-3.5-flash to quickly check this CSS file

Or programmatically:

{ "name": "ask-llm", "arguments": { "provider": "gemini", "model": "gemini-3.5-flash", "prompt": "..." } }

For Codex, common overrides:

Use ask-codex with model gpt-5.5-mini to summarize this commit

For Ollama, you can request any model you've pulled:

ollama pull deepseek-coder:6.7b

Use ask-ollama with model deepseek-coder:6.7b to review this implementation

Token Limits & Cost

Provider	Context window	Cost model
Gemini Pro	~1M tokens (~250k LOC)	Gemini Code Assist Standard/Enterprise seat (from 2026-06-18)
Gemini Flash	~1M tokens	Cheaper than Pro; fallback target for quota relief
Codex GPT-5.5	Per OpenAI's published context window	Per OpenAI billing
Codex GPT-5.5-mini	Smaller context	Cheaper; fallback target
Ollama	Per model (e.g., 256k for qwen3.6)	Free — runs locally

Track What You're Spending

Token usage is exposed live via:

• Per-call: result.structuredContent.usage on every ask-* tool response (provider, model, inputTokens, outputTokens, cachedTokens, thinkingTokens, durationMs, fellBack) — see ADR-054
• Per-session aggregate: call the get-usage-stats MCP tool, or read the usage://current-session MCP Resource for a JSON snapshot
• In the REPL: type /usage for a markdown-formatted breakdown

This is in-memory only — no persistence to disk, resets when the MCP server restarts.

Recommendations by Use Case

• General code review → defaults are correct; let the fallback chain handle quota
• Whole-codebase analysis → ask-gemini (Pro) if you have an enterprise seat, otherwise ask-antigravity for large-context reads without per-token billing
• Quick fixes, fast iteration → request Flash or gpt-5.5-mini explicitly to skip the Pro→fallback round-trip
• Privacy-sensitive code → ask-ollama, never leaves your machine
• Multi-perspective debate → multi-llm or /brainstorm skill — Claude weighs verified vs inferred