# NousResearch/hermes-agent · Context length override and 64K minimum conflicts for local models

## Open now

- 1 issue, 2 PRs open on GitHub.
- #60103: \[Bug\]: Compression unreachable on local models whose real context is silently clamped by Ollama — dead zone causes permanent context-shift and multi-minute latency
- PR #60154: fix: cap config context_length at real local provider window (#60103)
- PR #60460: fix(agent): use real model context when smaller than enforced 64K minimum for compression (#60103)

These members appear to center on one theme: how the configured or real model context length interacts with an enforced 64K minimum, particularly for local providers like Ollama where the actual window may be smaller. One is a bug report describing compression being unreachable, and the others are proposed fixes that honor overrides or cap to the real window.

## TL;DR

- 5 member(s): 1 open issue(s), 0 closed issue(s), 2 open PR(s), 0 merged PR(s), 2 PR(s) closed without merging.
- No fix has been merged to main; one or more PRs were attempted (see below).
- **Recommended:** review the prior/competing PR(s) below. A fix was attempted but not merged; your case may extend or revive that work.

## Maintainer resolution verdict

- Internal-tracker verdict over this cluster's issue members: `unresolved`. This is distinct from the fix status above. Weigh it yourself.

## How this cluster was grouped

- Link evidence (`STRUCTURAL_SPANNING`): every member is connected by structural evidence: shared changed files, or a closing/superseded reference.
- Pairwise link density `0.40`: the share of member 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.

## Open related issues

- #60103: \[Bug\]: Compression unreachable on local models whose real context is silently clamped by Ollama — dead zone causes permanent context-shift and multi-minute latency

## Prior & competing PRs

- **#8962: fix(agent): honor model.context_length override below 64K floor**. closed without merge
  - maintainer direction: category `wontfix_bydesign`; implication `wontfix`; refs `#53569`
- **#11097: fix(run_agent): honor model.context_length override for sub-64K models**. closed without merge
  - maintainer direction: category `wontfix_bydesign`; implication `wontfix`; refs `#53569`
- **#60154: fix: cap config context_length at real local provider window (#60103)**. open
  - maintainer direction: category `uncategorized`; implication `needs_llm`; refs `68a65ed7a`
  - files: `agent/model_metadata.py`
- **#60460: fix(agent): use real model context when smaller than enforced 64K minimum for compression (#60103)**. open
  - maintainer direction: category `uncategorized`; implication `needs_llm`
  - files: `agent/model_metadata.py`

## Fix status

- No merged fix on record.
- Subsystem: `agent`