feat(core): add support for Qwen3.5 and doubao-seed 2.0#2021
Merged
yuyutaotao merged 7 commits intomainfrom Feb 24, 2026
Merged
feat(core): add support for Qwen3.5 and doubao-seed 2.0#2021yuyutaotao merged 7 commits intomainfrom
yuyutaotao merged 7 commits intomainfrom
Conversation
- Add 'qwen3.5' to TVlModeValues type and VL_MODE_RAW_VALID_VALUES - Handle qwen3.5 in deepThink config (enable_thinking param) - Update en/zh docs: model-common-config and model-strategy https://claude.ai/code/session_014yyF69JPnGi5rRDRby3qxg
Deploying midscene with
|
| Latest commit: |
7b26ce2
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://ed35beb2.midscene.pages.dev |
| Branch Preview URL: | https://claude-add-qwen3-5-model-hre.midscene.pages.dev |
* feat: add doubao-seed model family for Doubao-Seed-2.0-Lite support Add 'doubao-seed' as a new model_family enum value that behaves identically to 'doubao-vision' for bbox adaptation, JSON preprocessing, deep thinking config, and empty content fallback. Update documentation in both English and Chinese to reflect the new model support. https://claude.ai/code/session_01FKpQMcv4q946MoPSJLvz9T * docs: recommend doubao-seed as primary MODEL_FAMILY, keep doubao-vision as compatible https://claude.ai/code/session_01FKpQMcv4q946MoPSJLvz9T --------- Co-authored-by: Claude <noreply@anthropic.com>
yuyutaotao
changed the title
quanru
pushed a commit
that referenced
this pull request
Feb 26, 2026
* feat: add qwen3.5 model family with same runtime behavior as qwen3-vl - Add 'qwen3.5' to TVlModeValues type and VL_MODE_RAW_VALID_VALUES - Handle qwen3.5 in deepThink config (enable_thinking param) - Update en/zh docs: model-common-config and model-strategy https://claude.ai/code/session_014yyF69JPnGi5rRDRby3qxg * fix: correct qwen3.5 model name to qwen3.5-plus (no "vl" suffix) https://claude.ai/code/session_014yyF69JPnGi5rRDRby3qxg * docs(site): sync Chinese model-strategy.mdx with English qwen3.5 changes * docs(site): qwen3.5 deepThink is enabled by default * docs(site): reorder qwen3.5 before qwen3-vl in deepThink defaults * feat(core): add support for Doubao-Seed-2.0-Lite model family (#2022) * feat: add doubao-seed model family for Doubao-Seed-2.0-Lite support Add 'doubao-seed' as a new model_family enum value that behaves identically to 'doubao-vision' for bbox adaptation, JSON preprocessing, deep thinking config, and empty content fallback. Update documentation in both English and Chinese to reflect the new model support. https://claude.ai/code/session_01FKpQMcv4q946MoPSJLvz9T * docs: recommend doubao-seed as primary MODEL_FAMILY, keep doubao-vision as compatible https://claude.ai/code/session_01FKpQMcv4q946MoPSJLvz9T --------- Co-authored-by: Claude <noreply@anthropic.com> * chore(core): fix lint --------- Co-authored-by: Claude <noreply@anthropic.com>
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
2 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
This PR adds support for Alibaba Cloud's Qwen3.5 model family to Midscene, treating it as a variant of Qwen3-VL with identical runtime behavior.
Key Changes
Documentation Updates
deepThinkoption documentation to include Qwen3.5 supportType System Updates
'qwen3.5'toTVlModeValuestype definition'qwen3.5'toVL_MODE_RAW_VALID_VALUESarrayRuntime Implementation
resolveDeepThinkConfig()to handle Qwen3.5 model family with the sameenable_thinkingparameter mapping as Qwen3-VLadaptBbox()coordinate system comment to include Qwen3.5 in the default normalized 0-1000 coordinate system groupImplementation Details
Qwen3.5 is implemented as a first-class model family that shares the same behavior as Qwen3-VL:
enable_thinkingparameter)The implementation treats Qwen3.5 as a peer to Qwen3-VL rather than a subtype, allowing for independent configuration and future divergence if needed.
https://claude.ai/code/session_014yyF69JPnGi5rRDRby3qxg