Repository navigation
fix(tencent): correct Hy3 input and output limits - #6126
Merged
Merged
Conversation
Contributor
|
The lab-model limit change is sound, but inheritance leaves one provider with ## Action items
- **[medium] [possible mistake]** `models/tencent/hy3.toml:16` - **Check:** After `base_model` deep-merge, provider `limit.input` must not exceed that host’s `limit.context`. **Why:** Adding `input = 192_000` on the shared Hy3 base is inherited by hosts that only override context. `providers/opencode/models/hy3-free.toml` keeps `context = 190_000` and does not set `input`, so the resolved model becomes input 192k > context 190k—the same class of bad budgeting metadata this PR aims to fix. **Action:** On that provider (and any other Hy3 host whose context is below 192k), set an explicit `limit.input` ≤ its context, or use `base_model_omit = ["limit.input"]` if the host should not expose a separate input ceiling. |
Contributor
Action items
|
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
limit.input = 192_000to the sharedtencent/hy3base model.limit.outputfrom64_000to128_000, retaining the existing256_000context window.The missing input ceiling lets clients using inherited metadata, including OpenCode Go, budget against the larger context window instead of the input limit. This change updates the shared base model; explicit provider overrides remain intact.
Related: anomalyco/opencode#45168, anomalyco/opencode#46137.
Source
Validation
bun validatepasses.repository open-weight model metadata includes weights links. All 39 models reported by that check are unchanged from the base commit.