Z.ai (GLM)

GLM models with an unusually cheap flat-rate coding plan

z.ai docs.z.ai

Overall in llm apis

15of 18 · score 58.4

Data coverage

78%

21 of 27 fields sourced

Best for

•Driving a coding agent all day on a predictable monthly cost
•Cheap high-volume code completion and refactoring
•Self-hosting an MIT-licensed coding model

Strong on

+Trains on your data Partial 2/18

Weak on

−Zero retention No 16/18
Better: Amazon Bedrock: Yes
−Context window 200,000 tokens 11/14
Better: Anthropic: 1,000,000 tokens
−Audio in/out No 11/18
Better: OpenAI: Partial

vs Cerebras

•Ahead on Trains on your data: Partial vs No
•Ahead on Effort control: Partial vs No
•Behind on Zero retention: No vs Yes
•Behind on $/M output: $2.2/M tok vs $0.75/M tok

How Z.ai (GLM) scores among 18 llm apis21 of 27 fields known

Capability7

Flagship model

GLM-4.6

· 2026-01-15

Context window

200,000 tokens

11/14

Leader: Anthropic · 1,000,000 tokens

· 2026-01-15

Max output

—

Image input

Partial

10/18

Leader: Anthropic · Yes

· 2026-01-15

Audio in/out

11/18

Leader: OpenAI · Partial

· 2026-01-15

Open weights1st

Yes

1/18

huggingface.co · 2026-01-15

OpenAI-compat API1st

Yes

1/18

· 2026-01-15

Performance2

TTFT p50

—

Output tok/s

—

Pricing4

$/M input

$0.6/M tok

6/13

Leader: Groq · $0.05/M tok

docs.z.ai · 2026-11-06

$/M output

$2.2/M tok

6/13

Leader: Fireworks AI · $0.05/M tok

docs.z.ai · 2026-11-06

$/M cache read

$0.11/M tok

4/12

Leader: DeepSeek · $0.0028/M tok

docs.z.ai · 2026-11-07

Batch discount

—

Agentic6

Tool use1st

Yes

1/18

· 2026-01-15

Schema output

Partial

11/18

Leader: Anthropic · Yes

· 2026-01-15

Effort control

Partial

5/18

Leader: Anthropic · Yes

· 2026-01-15

Computer use

5/18

Leader: Anthropic · Yes

· 2026-01-15

MCP support

6/18

Leader: Anthropic · Yes

· 2026-01-15

Cache TTL

—

Governance & continuity6

Continuity policy

Versioned model names (GLM-4.5, 4.6) stay addressable after a successor ships, but there is no published deprecation policy or notice period. MIT-licensed weights are the fallback: every flagship generation so far has been released publicly, so a retired endpoint does not strand the model.

· 2026-01-15

Notice period

—

Pinnable versions1st

Yes

1/18

· 2026-01-15

Zero retention

16/18

Leader: Amazon Bedrock · Yes

· 2026-01-15

Trains on your data2nd