Cairn

max_tokens silently dropped across providers and routes

NousResearch/hermes-agent · cluster 59763 · agent

Data as of

3 open items. Fix merged in #70279. Check it covers your case.

The short answer

2open issues
1open PR
10items grouped
Known?
Yes. 10 related items grouped here, 0 issues already closed.
Fix event
A fix was merged to main in PR #70279. Check it covers your case. No release data is tracked.
Tracker
The maintainers' tracker marks this unresolved. That is separate from the fix event.
Still open
2 issues and 1 PR.

A bug where max_tokens settings from config.yaml were silently ignored on various code paths, with multiple fixes forwarding the parameter to all providers and routes. One related item adds a reference_max_tokens cap for advisor output.

Open issues2

  • Issue#59763_resolve_runtime_agent_kwargs_for_provider()` omits `max_tokens` from its return dict, so `model.max_tokens` / `HERMES_MAX_TOKENS` / per-provider `max_output_tokens` caps are silently ignored on any provider-pinned route; `_build_call_kwargs()` additionally drops an explicit `max_tokens` for OpenRouter on the auxiliary path
  • Issue#60388[Bug]: `max_tokens` settings across config.yaml are silently dropped

Open PRs1

  • PR#59792fix(agent): include max_tokens in _resolve_runtime_agent_kwargs_for_provider return dict (#59763)

Fixes on record

Merged to main. A stamp counts the issues a fix is linked to.

  • #70279 fix(moa): reference_max_tokens caps advisors on every provider, never the aggregator + per-slot caps (cluster salvage)

    merged to main

    files: agent/auxiliary_client.py, agent/conversation_loop.py, agent/moa_loop.py, contributors/emails/[redacted:email], contributors/emails/[redacted:email], contributors/emails/[redacted:email], hermes_cli/moa_config.py, tests/agent/test_auxiliary_client.py, tests/agent/test_moa_context_max_tokens.py, tests/agent/test_moa_slot_max_tokens.py, tests/hermes_cli/test_moa_config.py

6 PRs closed without merging

From the close comments: closed without comment 2 · stale or abandoned 2 · other 2

  • #60454 fix(auxiliary): forward max_tokens to all providers, not just Anthropic/NIM (#60388) · stale or abandoned
  • #60513 fix(aux): forward max_tokens to all providers instead of dropping it (#60388) · stale or abandoned
  • #60620 fix: forward max_tokens to all providers in _build_call_kwargs (#60388) · closed without comment
  • #60645 fix(aux): forward max_tokens to all providers instead of dropping it for non-Anthropic/NIM · other → #60620
  • #71229 fix(gateway): forward max_tokens on provider-pinned routes (#59763) · closed without comment
Show 1 more
  • #124043 fix: resolve #60388 — [Bug]: max_tokens settings across config.yaml are silently dropped · other
How this cluster was grouped
  • Link evidence (partly structural links): some members are connected by structural evidence (shared changed files, or a closing/superseded reference); the rest are joined by an automated relevance judgement over their text.
  • Members outside that structural core may belong to a different underlying problem.
  • Pairwise link density 0.19: the share of item pairs that carry a link, counting a structural link in full and a relevance-only link at 0.36. 1.00 means every pair is linked.

Pick up a Lab task

Check whether a fix covers the open issues here, with your own agent. Bring back evidence for maintainers.

Open the Lab

At a glance

Subsystem
agent
Merged PRs
#70279
Closed issues
0
Attempts
6 PRs closed without merging
Fix link checked
23 Jul 2026

For your agent

The same facts, typed.

cluster/59763.json
cluster/59763.md