GLM
Z.ai's frontier line — reasoning always on, a 1M context.
GLM is Z.ai's main line. GLM-5.3 turns reasoning on always (effort low to max, max by default), reads a 1M-token context, and made its largest coding jump in the family's history over GLM-5.2, on Z.ai's own benchmarks. Older versions stay served at lower prices.
What it is good at
- Agentic coding — Terminal-Bench 3.0 28.3 against GLM-5.2's 4.6, DeepSWE v1.1 66.9 against 46.2 (vendor-reported)
- Reasoning always on with adjustable effort; a 1M context at $1.40 in / $4.40 out per MTok
- Older versions still served at lower prices (GLM-5, then 4.7/4.6/4.5)
What it is less good at
- Text-only input — no vision
- The benchmarks are Z.ai's own, measured its way; no independent figures may be republished here
Reaching it with TeamCLI
TeamCLI runs GLM as GLM 5.3:
Ways to reach it
| Route | Spends | Status |
|---|---|---|
| Claude Code with a provider endpoint | A Z.ai (Zhipu) API key or coding plan. | Verified |
| The API directly | Z.ai's API billing. | Listed — not yet verified |
The seat line
Each route maps to a TeamCLI seat line — the same fields, written generically. No machine paths, no launcher script of anybody's, no team's own seat names.
| Route | cli | vendor | model | version | display | launch |
|---|---|---|---|---|---|---|
| Claude Code with a provider endpoint | claude-code | zhipu | GLM | 5.3 | GLM 5.3 | my-lab-claude |
A seat with model_from: launcher leaves the running model to what the launch runs — the provider's launcher, or Jules — so team doctor reads the model from the seat rather than from the file.
Known quirks
- Z.ai GLM ships no seat CLI of its own; it provides a coding helper that configures other CLIs, and the verified route here is Claude Code with its endpoint.
Versions
One block per version, the current one first, each with its own anchor. The comparison table puts the costs and the family's benchmark columns side by side; the values are verbatim, and every one is sourced and dated in its version's section below and in the Sources list. A dash means the source publishes no figure, and nothing here is estimated.
No independent intelligence or speed index is republished here: Artificial Analysis measures both across labs, and its terms allow the link but not the copy — its figures live on its own pages.
| Version | Released | Context | Input price | Output price | Z.ai Code Bench (max effort) | Terminal-Bench 3.0 (agentic coding) | DeepSWE v1.1 (agentic coding) | Agents' Last Exam |
|---|---|---|---|---|---|---|---|---|
| GLM-5.3 — current | — | 1M tokens | $1.40 / MTok | $4.40 / MTok | 34.5% at ~75K output tokens | 28.3 | 66.9 | 28.5 |
| GLM-5.2 | — | — | $1.40 / MTok | $4.40 / MTok | 23.4% | 4.6 | 46.2 | 23.8 |
| GLM-5.1 | — | — | $1.40 / MTok | $4.40 / MTok | — | — | — | — |
| GLM-5 | — | — | $1.00 / MTok | $3.20 / MTok | — | — | — | — |
| GLM-4.7 | — | — | $0.60 / MTok | $2.20 / MTok | — | — | — | — |
| GLM-4.6 | — | — | $0.60 / MTok | $2.20 / MTok | — | — | — | — |
| GLM-4.5 | — | — | $0.60 / MTok | $2.20 / MTok | — | — | — | — |
GLM-5.3 — current
- Model id:
glm-5.3
Abilities
- Released: —
- Context window: 1M tokens — Z.ai — GLM-5.3 model page (read 2026-10-19).
- Max output: 128K tokens — Z.ai — GLM-5.3 model page (read 2026-10-19).
- Output speed: —
- Latency: —
Costs
- Input price: $1.40 / MTok — Z.ai — pricing (read 2026-10-19). Cached input $0.26; text-only input; reasoning always on (low/high/max, default max)
- Output price: $4.40 / MTok — Z.ai — pricing (read 2026-10-19).
Benchmarks — as reported by the lab
- Z.ai Code Bench (max effort) — 34.5% at ~75K output tokens (Z.ai — GLM-5.3 model page, read 2026-10-19)
- Terminal-Bench 3.0 (agentic coding) — 28.3 (Z.ai — GLM-5.3 model page, read 2026-10-19)
- DeepSWE v1.1 (agentic coding) — 66.9 (Z.ai — GLM-5.3 model page, read 2026-10-19)
- Agents' Last Exam — 28.5 (Z.ai — GLM-5.3 model page, read 2026-10-19)
This is the version the seat lines above run.
GLM-5.2
- Model id:
glm-5.2
Abilities
- Released: —
- Context window: —
- Max output: —
- Output speed: —
- Latency: —
Costs
- Input price: $1.40 / MTok — Z.ai — pricing (read 2026-10-19). GLM-5.3's base model; 5.3 is post-training on it
- Output price: $4.40 / MTok — Z.ai — pricing (read 2026-10-19).
Benchmarks — as reported by the lab
- Z.ai Code Bench (max effort) — 23.4% (Z.ai — GLM-5.3 model page, read 2026-10-19)
- Terminal-Bench 3.0 (agentic coding) — 4.6 (Z.ai — GLM-5.3 model page, read 2026-10-19)
- DeepSWE v1.1 (agentic coding) — 46.2 (Z.ai — GLM-5.3 model page, read 2026-10-19)
- Agents' Last Exam — 23.8 (Z.ai — GLM-5.3 model page, read 2026-10-19)
GLM-5.1
- Model id: —
Abilities
- Released: —
- Context window: —
- Max output: —
- Output speed: —
- Latency: —
Costs
- Input price: $1.40 / MTok — Z.ai — pricing (read 2026-10-19).
- Output price: $4.40 / MTok — Z.ai — pricing (read 2026-10-19).
GLM-5
- Model id: —
Abilities
- Released: —
- Context window: —
- Max output: —
- Output speed: —
- Latency: —
Costs
- Input price: $1.00 / MTok — Z.ai — pricing (read 2026-10-19).
- Output price: $3.20 / MTok — Z.ai — pricing (read 2026-10-19).
GLM-4.7
- Model id: —
Abilities
- Released: —
- Context window: —
- Max output: —
- Output speed: —
- Latency: —
Costs
- Input price: $0.60 / MTok — Z.ai — pricing (read 2026-10-19).
- Output price: $2.20 / MTok — Z.ai — pricing (read 2026-10-19).
GLM-4.6
- Model id: —
Abilities
- Released: —
- Context window: —
- Max output: —
- Output speed: —
- Latency: —
Costs
- Input price: $0.60 / MTok — Z.ai — pricing (read 2026-10-19).
- Output price: $2.20 / MTok — Z.ai — pricing (read 2026-10-19).
GLM-4.5
- Model id: —
Abilities
- Released: —
- Context window: —
- Max output: —
- Output speed: —
- Latency: —
Costs
- Input price: $0.60 / MTok — Z.ai — pricing (read 2026-10-19).
- Output price: $2.20 / MTok — Z.ai — pricing (read 2026-10-19).
Sources
- Z.ai — GLM-5.3 model page — read 2026-10-19
- Z.ai — pricing — read 2026-10-19
Figures refreshed 2026-10-19; benchmarks are vendor-reported. Prices are per million tokens unless a figure says otherwise.