Amp Model Override Cheat Sheet

How to Override the Default Model

Option 1: amp.model.sonnet (simple toggle)

{
  "amp.model.sonnet": true
}

This specifically switches Smart mode from Gemini 3 Pro Preview to Claude Sonnet 4.5.

Option 2: amp.internal.model (flexible override)

{
  "amp.internal.model": "anthropic:claude-sonnet-4-5-20250929"
}

This lets you pick any model using provider:model-name format.

Note: internal.model takes precedence if both settings are configured.

---

All Available Models

ANTHROPIC

Setting Value	Display Name	Context	Max Output
anthropic:claude-sonnet-4-5-20250929	Claude Sonnet 4.5	200K	64K
anthropic:claude-sonnet-4-20250514	Claude Sonnet 4	200K	32K
anthropic:claude-opus-4-1-20250805	Claude Opus 4.1	200K	32K
anthropic:claude-opus-4-20250514	Claude Opus 4	200K	32K
anthropic:claude-haiku-4-5-20251001	Claude Haiku 4.5	200K	64K
anthropic:claude-3-5-sonnet-20241022	Claude 3.5 Sonnet	200K	8K
anthropic:claude-3-5-haiku-20241022	Claude 3.5 Haiku	200K	8K

VERTEXAI (Google)

Setting Value	Display Name	Context	Max Output
vertexai:gemini-3-pro-preview	Gemini 3 Pro Preview	1M	65K
vertexai:gemini-2.5-pro	Gemini 2.5 Pro	1M	65K
vertexai:gemini-2.5-flash	Gemini 2.5 Flash	1M	65K
vertexai:gemini-2.5-flash-lite	Gemini 2.5 Flash Lite	1M	65K
vertexai:gemini-2.5-flash-preview-09-2025	Gemini 2.5 Flash (Preview)	1M	65K
vertexai:gemini-2.5-flash-lite-preview-09-2025	Gemini 2.5 Flash Lite (Preview)	1M	65K

OPENAI

Note: Amp uses an internal registry (Wm4) for OpenAI model routing. Models not in this registry will silently fall back to gpt-5.1 due to prefix matching.

Available models:

Setting Value	Display Name	Context	Max Output
openai:gpt-5.1	GPT-5.1	400K	128K
openai:gpt-5-codex	GPT-5 Codex	400K	128K
openai:gpt-5-mini	GPT-5 Mini	400K	128K
openai:gpt-5-nano	GPT-5 Nano	400K	128K
openai:gpt-4.1	GPT-4.1	1M	32K
openai:gpt-4.1-mini	GPT-4.1 Mini	-	-
openai:gpt-4.1-nano	GPT-4.1 Nano	-	-
openai:gpt-4o	GPT-4o	128K	16K
openai:gpt-4o-mini	GPT-4o Mini	-	-
openai:o3	o3	200K	-
openai:o3-mini	o3-mini	200K	-
openai:o3-deep-research	o3 Deep Research	-	-
openai:o4-mini-deep-research	o4 Mini Deep Research	-	-

Not available (will fall back to gpt-5.1):

Setting Value	Fallback
openai:gpt-5.1-codex	gpt-5.1
openai:openai/gpt-oss-120b	gpt-5.1

XAI

Setting Value	Display Name	Context	Max Output
xai:grok-code	Grok Code	256K	32K
xai:grok-code-fast-1	Grok Code Fast 1	256K	32K

MOONSHOT

Setting Value	Display Name	Context	Max Output
moonshotai:kimi-k2-instruct-0905	Kimi K2 Instruct	-	-

OPENROUTER

Note: OpenRouter supports passthrough to any model available on their platform. The model ID after openrouter: is passed directly to OpenRouter's API.

Pre-defined model:

Setting Value	Display Name	Context	Max Output
openrouter:sonoma-sky-alpha	Sonoma Sky Alpha	256K	32K

Passthrough to any OpenRouter model:

You can use any model available on OpenRouter by specifying its full model ID:

{
  "amp.internal.model": "openrouter:anthropic/claude-3-opus-20240229"
}

Examples:

• openrouter:anthropic/claude-3-opus-20240229
• openrouter:meta-llama/llama-3.1-405b-instruct
• openrouter:google/gemini-pro-1.5
• openrouter:mistralai/mixtral-8x22b-instruct

Requires: amp.openrouter.apiKey setting or OPENROUTER_API_KEY environment variable.

CEREBRAS

Setting Value	Display Name	Context	Max Output
cerebras:zai-glm-4.6	Z.ai GLM 4.6	131K	40K

---

Mode Defaults

Mode	Default Model
Smart	vertexai:gemini-3-pro-preview
Free	anthropic:claude-haiku-4-5-20251001
Dojo	anthropic:claude-sonnet-4-5-20250929
Dojo (Free)	anthropic:claude-haiku-4-5-20251001