NousResearch/hermes-agent

GLM-5.3 graded reasoning_effort support fixes

Cluster 91789 · agent/transports. Data as of 2026-10-25 11:21 UTC

Nothing open on GitHub. Fix merged in #91844 on 21 Aug 2026. Check it covers your case.

open issues

open PRs

items grouped

Known?: Yes. 4 related items grouped here, 1 issue already closed.
Fix event: A fix was merged to main in PR #91844. Check it covers your case. No release data is tracked.
Tracker: The maintainers' tracker marks this resolved, deciding PR #91844. That is separate from the fix event.
Still open: Nothing open on GitHub.

Next step: read #91844, then check it covers your case.

Read #91844 Pick up a Lab task

These issues and pull requests address a bug where GLM-5.3 was incorrectly handled by a version-check function meant for GLM-5.2, causing the provider to send only binary thinking signals instead of the graded reasoning_effort values (low/medium/high/max) that GLM-5.3 supports.

Open issues0

No open issues.

Open PRs0

No open pull requests.

Fixes on record

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

#91844 fix: GLM-5.3 low/medium reasoning effort no longer clamps to high (#91789)

merged to main

files: agent/reasoning_effort.py, plugins/model-providers/zai/__init__.py, tests/plugins/model_providers/test_zai_profile.py

2 PRs closed without merging

From the close comments: done elsewhere 1 · superseded by another PR 1

#91840 fix(zai): send graded reasoning_effort for glm-5.3, not just binary thinking · superseded by another PR → #91844, #91789
#91965 fix(opencode): emit graded reasoning_effort for glm-5.3 on opencode-go · done elsewhere → #91844

Closed issues · 1

Closed#91789[Bug] zai: glm-5.3 accepts graded reasoning_effort (low/medium/high/max) but provider plugin only sends binary thinking — _is_glm_5_2() misses glm-5.3

How this cluster was grouped

Link evidence (structural links throughout): every member is connected by structural evidence: shared changed files, or a closing/superseded reference.
Pairwise link density 0.50: 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/transports
Merged PRs: #91844
Closed issues: 1
Attempts: 2 PRs closed without merging
Fix link checked: 21 Aug 2026

For your agent

The same facts, typed.

cluster/91789.json
cluster/91789.md