Claude Code Models Catalog

Claude Code Models Catalog

Claude Code has a built-in list of 22 models. The list tells Claude Code the name, the context window, and the output limit of each model. To get a 1M window, find native 1M in the table, and then run claude --model opus.

About the Catalog

The catalog is a JavaScript object in the bundle that carries the comment Hand-maintained baked-in model catalog. It holds the provider names, the window, the output limits, the price, and the capability flags of each model. A row has the mark ✅ when a page of the Model Configuration docs or another docs page names the model. It has the mark 🔍 when no docs page names it. The docs do not list the catalog fields, so the window, the limits, and the prices are bundle facts for all rows.

Catalog

The column Window is the context window of the model. The column Larger window says how to get a bigger one: native 1M means the model has a 1M window already, the context-1m beta asks the API for it, and the [1m] suffix asks for it in --model. The column Output shows the default and the highest output limit in tokens. The column Price shows the dollars for one million input tokens and one million output tokens. See Models Capabilities for the cache prices.

id	Mark	Name	Knowledge cutoff	Window	Larger window	Output	Default effort	Price
claude-3-5-haiku	🔍	Haiku 3.5	-	not set	-	8,192 / 8,192	-	$0.8 / $4
claude-haiku-4-5	✅	Haiku 4.5	February 2025	200,000	[1m] suffix	32,000 / 64,000	-	$1 / $5
claude-haiku-5-5	✅	Haiku 5.5	June 2026	1,000,000	native 1M, context-1m beta	128,000 / 128,000	medium	$0.1 / $0.5
claude-3-5-sonnet	🔍	Sonnet 3.5	-	not set	-	8,192 / 8,192	-	$3 / $15
claude-3-7-sonnet	🔍	Sonnet 3.7	-	not set	-	32,000 / 64,000	-	$3 / $15
claude-sonnet-4-0	🔍	Sonnet 4	January 2025	200,000	context-1m beta, [1m] suffix	32,000 / 64,000	-	$3 / $15
claude-sonnet-4-5	✅	Sonnet 4.5	January 2025	200,000	context-1m beta, [1m] suffix	32,000 / 64,000	-	$3 / $15
claude-sonnet-4-6	✅	Sonnet 4.6	August 2025	200,000	context-1m beta, [1m] suffix	32,000 / 128,000	-	$3 / $15
claude-sonnet-5	✅	Sonnet 5	January 2026	1,000,000	native 1M, context-1m beta	64,000 / 128,000	high	$2 / $10
claude-sonnet-5-5	✅	Sonnet 5.5	June 2026	1,000,000	native 1M, context-1m beta	128,000 / 128,000	medium	$2 / $10
claude-opus-4-0	🔍	Opus 4	January 2025	200,000	[1m] suffix	32,000 / 32,000	-	$15 / $75
claude-opus-4-1	✅	Opus 4.1	January 2025	200,000	[1m] suffix	32,000 / 32,000	-	$15 / $75
claude-opus-4-5	✅	Opus 4.5	May 2025	200,000	[1m] suffix	32,000 / 64,000	-	$5 / $25
claude-opus-4-6	✅	Opus 4.6	May 2025	200,000	context-1m beta, [1m] suffix	64,000 / 128,000	-	$5 / $25
claude-opus-4-7	✅	Opus 4.7	January 2026	1,000,000	native 1M, context-1m beta, [1m] suffix	64,000 / 128,000	xhigh	$5 / $25
claude-opus-4-8	✅	Opus 4.8	January 2026	1,000,000	native 1M, context-1m beta, [1m] suffix	64,000 / 128,000	high	$5 / $25
claude-opus-5	✅	Opus 5	May 2026	1,000,000	native 1M, context-1m beta, [1m] suffix	64,000 / 128,000	high	$5 / $25
claude-opus-5-5	✅	Opus 5.5	June 2026	1,000,000	native 1M, context-1m beta, [1m] suffix	128,000 / 128,000	medium	$4 / $20
claude-fable-5	✅	Fable 5	January 2026	1,000,000	native 1M, context-1m beta	64,000 / 128,000	high	$10 / $50
claude-fable-5-1	✅	Fable 5.1	June 2026	1,000,000	native 1M, context-1m beta	64,000 / 128,000	high	$10 / $50
claude-mythos-5	✅	Mythos 5	January 2026	1,000,000	native 1M, context-1m beta	64,000 / 128,000	-	$10 / $50
claude-mythos-5-1	🔍	Mythos 5.1	June 2026	1,000,000	native 1M, context-1m beta	64,000 / 128,000	high	$10 / $50

Notes

• A model with no window in the table has no context entry in the catalog.
• The catalog keeps a rank for each model for the advisor tool (--advisor). The ranks run from 1 for claude-haiku-4-5 to 9 for claude-fable-5-1 and claude-mythos-5-1.
• A model can name a model to use instead on a third-party provider. For example, claude-opus-5-5 falls back to claude-opus-5, and claude-fable-5-1 falls back to claude-fable-5.
• The Mythos models have a name on the Anthropic API only, and none on the cloud providers.
• The lean_prompt capability chooses the short system prompt for a model (see System Prompt).

See Also

• Models and API
• Models Providers
• Models Capabilities