Agent re-calls a reasoning model until max_agent_steps when every reply is empty because reasoning exhausts max_output_tokens

With the default text exit condition, an Agent treated an empty reply caused by the model spending its whole output budget on reasoning as an invalid reply and retried. Because the truncation was deterministic, every retry was empty and the Agent kept calling the model until max_agent_steps, wasting all reasoning tokens.

Framework / agent

Haystack · tool-calling agent loop

Remediation attempts

TESTED

Recurrence

not documented

Source languages

en

Updated

2026-09-29

Sources

• GITHUB ISSUE `Agent` loops to `max_agent_steps` when a reasoning model returns no text because it hits `max_output_tokens` — github.com/deepset-ai/haystack, retrieved 2026-09-29
• GITHUB PULL REQUEST fix: stop Agent retries after output limit — github.com/deepset-ai/haystack, retrieved 2026-09-29

Symptoms

• The Agent keeps calling the LLM until max_agent_steps, receiving an empty assistant reply each time
• Reasoning tokens are burned on every wasted call (1280 across a 5-step run in the report)

Similarity to your system is not implied. A remediation that worked in the documented context may not work in yours.