Long Decode Continuation
A service feature designed by Google for extremely long text generation scenarios. It addresses request timeout interruptions by pausing mid-generation during lengthy responses and resuming in subsequent calls, supporting continuous output at the million-token scale.
Timeline (last 90 days)
Gemini 4 Argon Long Decode Continuation requires setting background=True and store=true, with max_output_tokens higher than 65,536, advancing in hops of at most 64K tokens, and running for up to 24 hours.
Gemini 4 Argon achieves 1 million token output through the Gemini API's new Long Decode Continuation feature, which pauses long responses and resumes them in subsequent calls to avoid timeouts.
Google has not yet published the request parameters or response fields of Long Decode Continuation.
Argon relies on a service feature called Long Decode Continuation, which pauses during an extremely long response and resumes generation in a subsequent call.
All Facts (4)
Gemini 4 Argon Long Decode Continuation requires setting background=True and store=true, with max_output_tokens higher than 65,536, advancing in hops of at most 64K tokens, and running for up to 24 hours.
50%UnverifiedGemini 4 Argon achieves 1 million token output through the Gemini API's new Long Decode Continuation feature, which pauses long responses and resumes them in subsequent calls to avoid timeouts.
50%UnverifiedGoogle has not yet published the request parameters or response fields of Long Decode Continuation.
50%UnverifiedArgon relies on a service feature called Long Decode Continuation, which pauses during an extremely long response and resumes generation in a subsequent call.
50%