Z.ai (Zhipu AI)
Z.ai's official GLM-5.2 harness: a desktop agentic development environment to plan, code, review, and deploy, with a multi-agent workspace, subagents, Goal Mode, skills, MCP servers, a remote/cloud workspace, and an embedded headless CLI inside the app bundle.
Attributed output
GitHub public-attribution count, 2026-10-11
1,253Attributed outputGitHub public-attribution count, 2026-10-11Commits and PRs on public GitHub provably produced by this agent. Exact if it is visible through its own bot or workflow, an estimate if measured through attribution it declares (these can swing 30 to 40 percent between measurement runs at multi-million scale), and a floor either way.Index rank · all agents
#9 of 40 CLI agents
#10Index rank · all agents#9 of 40 CLI agentsRanked by agent-opened PRs in the trailing 4 complete weeks, matching the index. Agents without a tracked PR series follow, ordered by total attributed output. Other filters or sorts change the list position.n/a
Business signal
No sourced business metric
n/aBusiness signalNo sourced business metricn/a
Entry price
No published pricing
n/aEntry priceNo published pricingNot published
GitHub stars
No public repo
Not publishedGitHub starsNo public repoNot published
npm demand/wk
Public npm counter
Not publishednpm demand/wkPublic npm counterWeekly package downloads as an install-demand floor; adjusted when the vendor's main install channel is no longer npm.From the public record
359 PRs · 142 repositories
Search window: – · captured Oct 14, 2026 (UTC)
JustVugg/colibri: “sampling: partial top-keep select in attention_rows DSA — O(nk) quickselect, not O(nk log nk) qsort (#356)” (merged). storytold/wordcraft: “Add Brazilian Portuguese (pt-br) interface translation” (merged).
This is a sample of publicly attributed work, not total usage or a market-share estimate. Repositories are not customers. Attribution is declared in metadata, not independently verified tool execution; it does not prove exclusive authorship or autonomy. PR status is recorded at capture; merged does not establish correctness.
Up to 100 most recently created search results in each of five consecutive 28-day windows; post-filtered for attribution in the PR body, bot author, or Codex branch. Not a random or exhaustive sample.
Returned records dated Jun 13, 2026–Oct 11, 2026.
Download the dated audit evidence (JSON)Weekly attributed output on public GitHub. Output visible through the agent's own bot or workflow is exact; output measured through declared attribution is an estimate. Window: 2026-06-09 to 2026-10-05.
Public GitHub PRs attributed to ZCode
Derived from ZCode's tracked public GitHub series. A floor: work done under the developer's own identity, squash-merges, and private repositories are invisible.
Last full week (ending 2026-10-11): 173 agent-marked PRs (+29.1% week over week), search estimate.
Distinct users: 113 humans opened the 417 most recently sampled PRs, about 3.7 PRs per person.
Merge rate: 91.7% of its pull requests get merged (682 of 744 over 8 settled weeks, 2026-08-10 to 2026-10-04). The window ends about ten days back so recently-opened PRs have had time to merge.
GitHub review activity: 2.5% of its merged PRs had at least one GitHub review record before merging; the other 97.5% had none. This is a broad weekly signal and can include bot reviews; the human-review read below comes from the enriched PR sample.
In significant repos (company-owned or 10+ stars), 31.6% of its merged PRs got a human review. Across all sampled repos, the rate is 25%. Human review means a non-author, non-bot GitHub reviewer in the PR-lifecycle sample.
Median time to merge: 20 min. The unreviewed median is 0.19h. The human-reviewed median is 1.04h. From a 120-PR sample merged 2026-07-30 to 2026-08-28.
Test-inclusion: 60.8% of its sampled merged PRs touch a test file. This is a rough read on production readiness. 7 large PRs hit GitHub's 100-file path read cap, so treat this as a floor.
Snapshot 2026-10-11, confidence Medium, tracked since 2026-06-09, +16.2% over 7 days. Read the methodology.
| Signal | Value | Type | Confidence |
|---|---|---|---|
| Latest release | v3.2.4 (2026-07-03) | Official | High |
| Flagship model | GLM-5.2 — Terminal-Bench 2.1 81.0, SWE-bench Pro 62.1 (strongest open-source coding model) | Official | High |
| GLM-5 open weights | 6,129 GitHub stars (Apache-2.0) | Public metric | High |
Share of ZCode's own classified sampled work by segment; countries reflect where its PR authors are based. Full breakdowns: Languages, Frameworks, Industries, Map.
Summary checked 2026-07-03.
The ZCode desktop app is a free download; model usage runs on the GLM Coding Plan subscription (Lite / Pro / Max), with GLM Coding Max at $160/month ($1,344/year). The same plan can also be pointed at Claude Code, Cline, and OpenCode.