Is qwen3-max-thinking deprecated?

Q: Is qwen3-max-thinking deprecated?

No — still supported. Listed by litellm, openrouter. No retirement date announced. Neither the provider's API nor its documentation has been checked for this model.

NOT DEPRECATED

Listed by litellm, openrouter. No retirement date announced. Neither the provider's API nor its documentation has been checked for this model. Sources disagree on: context_tokens, input_price_per_mtok, output_price_per_mtok.

state=alive sunset=none days_remaining=n/a severity=low sources=litellm,openrouter evidence_tier=catalog confirmed_by_endpoint=false documented_by_provider=false replacement=none corroborated=yes divergent_fields=context_tokens,input_price_per_mtok,output_price_per_mtok checked=2026-09-16 03:02Z

alibaba/qwen3-max-thinking

Vendor: alibaba. Tracked by 2 of 5 sources.

What each source reports

**endpoint** = the provider's own machine-readable API, the strongest evidence. **provider docs** = the provider's own documentation, read by a model — first-party content, inferred reading. Unlabelled rows are third-party catalogs. Where sources disagree both values are shown rather than silently reconciled.
Source	ID	Context	Max out	In $/Mtok	Out $/Mtok	Sunset
litellm	deepinfra/Qwen/Qwen3-Max-Thinking	256,000	65,536	$1.2	$6	—
openrouter	qwen/qwen3-max-thinking	262,144	65,536	$0.78	$3.9	—

Change history

Severity	Change	What	Detected
low	added	new model openrouter/qwen/qwen3-max-thinking (262,144 ctx)	2026-09-06
low	added	new model deepinfra/Qwen/Qwen3-Max-Thinking (256,000 ctx)	2026-08-28