Skip to content

feat(catalog): show 8K/1K for GLM-5.2 / feat(catalog):为 GLM-5.2 显示 8K/1K - #1781

Merged
Oseltamivir merged 2 commits into
mainfrom
feat/glm52-8k1k-scenario
Nov 14, 2026
Merged

Oseltamivir merged 2 commits into
mainfrom
feat/glm52-8k1k-scenario

Conversation

@Oseltamivir

@Oseltamivir Oseltamivir commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

GLM-5.2 currently shows only AgentX. 8k1k is coming back for it in InferenceX:

Changes:

  • data-mappings.ts: removes Sequence.EightK_OneK from GLM-5.2's MODEL_HIDDEN_SEQUENCES and MODEL_DEPRECATED_SEQUENCES entries, so the 8K/1K tab shows with the default category.
  • data-mappings.test.ts: the two old GLM-5.2 tests (hidden 8K/1K, deprecated category) are replaced by one asserting 8K/1K is visible and default.

Ingest needs no change: rows resolve by model_prefix glm5.2 before the model-id map.

中文

GLM-5.2 目前只显示 AgentX。InferenceX 正在为它重新提供 8k1k:

改动:

  • data-mappings.ts:从 GLM-5.2 的 MODEL_HIDDEN_SEQUENCES 和 MODEL_DEPRECATED_SEQUENCES 条目中移除 Sequence.EightK_OneK,8K/1K 标签页以默认类别显示。
  • data-mappings.test.ts:原先两个 GLM-5.2 测试(隐藏 8K/1K、已弃用类别)替换为一个测试,断言 8K/1K 可见且类别为 default。

数据导入无需改动:数据行会先按 model_prefix glm5.2 解析,再查模型 ID 映射。

🤖 Generated with Claude Code


Note

Low Risk
Catalog visibility and deprecation flags only; no auth, ingest, or data-path changes.

Overview
GLM-5.2 again exposes the 8K/1K sequence in the catalog and options API, aligned with new InferenceX coverage for that model.

data-mappings.ts drops Sequence.EightK_OneK from GLM-5.2 in both MODEL_HIDDEN_SEQUENCES and MODEL_DEPRECATED_SEQUENCES, so 8K/1K is visible and categorized as default (same pattern as GLM-5.3-Flash). Tests for catalog mappings, GET /api/v1/views/options, and the video-history projection source pin are updated to match.

Reviewed by Cursor Bugbot for commit 3b46251. Bugbot is set up for automated code reviews on this repo. Configure here.

GLM-5.2 8k1k is back in InferenceX (B200 NVFP4 SGLang #5077, H200 FP8
SGLang #5078, H100 NVFP4 vLLM #5079), so stop hiding the sequence and
stop marking it deprecated for this model.

中文:feat(catalog):为 GLM-5.2 显示 8K/1K。InferenceX 重新为 GLM-5.2 提供 8k1k(B200 NVFP4 SGLang #5077、H200 FP8 SGLang #5078、H100 NVFP4 vLLM #5079),因此不再隐藏该序列,也不再将其标记为已弃用。
@vercel

vercel Bot commented Nov 14, 2026 •

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
inferencemax-app Ready Ready Preview Nov 14, 2026 11:23am UTC

Request Review

@Oseltamivir
Oseltamivir merged commit e713065 into main Nov 14, 2026
38 checks passed
@Oseltamivir
Oseltamivir deleted the feat/glm52-8k1k-scenario branch November 14, 2026 12:37

This branch was successfully deployed

1 active deployment
Preview — 3b462517 Deployed Nov 14, 2026 by vercel[bot]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant