fix(metadata): qwen3.6-plus has a 1M context window (#27008)#27068
Closed
haran2001 wants to merge 1 commit into
Closed
fix(metadata): qwen3.6-plus has a 1M context window (#27008)#27068haran2001 wants to merge 1 commit into
haran2001 wants to merge 1 commit into
Conversation
qwen3.6-plus did not have an explicit entry in DEFAULT_CONTEXT_LENGTHS, so the longest-substring fallback matched the generic 'qwen': 131072 catch-all. That dropped the effective context limit from 1,048,576 tokens to 131,072, prematurely lowered the compression threshold, and produced misleading warnings about main/compression context mismatch in long sessions. Add an explicit 'qwen3.6-plus': 1048576 entry before the catch-all and cover it with a regression test (bare, qwen/, and dashscope/ prefixes). Note: PR NousResearch#6599 also mentions touching model_metadata.py but the actual diff only edits hermes_cli/models.py, so this fix is independent and not duplicated by that PR. Closes NousResearch#27008
This comment was marked as spam.
This comment was marked as spam.
2 tasks
teknium1
added a commit
that referenced
this pull request
May 17, 2026
…tors Adds release-note attribution mappings for the contributors from group 5: - @haran2001 (PR #27070, #27068) - @ms-alan (PR #26443) - @godlin-gh (PR #26118) - @wesleysimplicio (PR #25777, ext-email form) - @Carry00 (PR #26851) - @alaamohanad169-ship-it (PR #26036) - @hawknewton (PR #26294) (YanzhongSu PR #25879 and flamiinngo PR #27231 already mapped.)
Contributor
|
Merged via PR #27382 — your commit was cherry-picked onto current |
gweeteve
pushed a commit
to gweeteve/hermes-agent
that referenced
this pull request
Jun 2, 2026
…tors Adds release-note attribution mappings for the contributors from group 5: - @haran2001 (PR NousResearch#27070, NousResearch#27068) - @ms-alan (PR NousResearch#26443) - @godlin-gh (PR NousResearch#26118) - @wesleysimplicio (PR NousResearch#25777, ext-email form) - @Carry00 (PR NousResearch#26851) - @alaamohanad169-ship-it (PR NousResearch#26036) - @hawknewton (PR NousResearch#26294) (YanzhongSu PR NousResearch#25879 and flamiinngo PR NousResearch#27231 already mapped.)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes #27008.
qwen3.6-pluswas missing fromDEFAULT_CONTEXT_LENGTHSinagent/model_metadata.py. The longest-substring fallback then matched the generic"qwen": 131072catch-all, dropping the effective context limit from 1,048,576 → 131,072 tokens, prematurely lowering the compression threshold and producing misleading "compression model context lower than main model" warnings during long sessions on DashScope / Alibaba / OpenRouter.Fix
Add an explicit entry before the generic catch-all:
The lookup is longest-key-first substring matching, so this works for bare
qwen3.6-plusand all provider-prefixed variants (qwen/qwen3.6-plus,dashscope/qwen3.6-plus, …).Not a duplicate of #6599
PR #6599's description mentions touching
agent/model_metadata.pybut its actual diff only adds the model tohermes_cli/models.py(the curated model list). The metadata table is independent and still defaults to 128K without this PR.Tests
New
test_qwen3_6_plus_context_lengthasserts 1,048,576 for bare and both provider-prefixed variants. Existingtest_qwen_generic_context_length(asserting 131072 for unknownqwen*IDs) still passes — no regression.