Compare models

Context, limits, capabilities, price and use cases side by side — up to 4 models at a time. The URL holds your selection, so you can share it.

GLM-5.1GLM-5.2GLM-5 Turbo Add a modelChat with these 3

AttributeHide identical rows	GLM-5.1z-ai/glm-5.1	GLM-5.2z-ai/glm-5.2	GLM-5 Turboz-ai/glm-5-turbo
Specs
Context window	200K	1M	200K
Max output	128K	128K	128K
Input modalities	Text	Text	Text
Output modalities	Text	Text	Text
Capabilities
Streaming	Supported	Supported	Supported
Tool calling	Supported	Supported	Supported
Structured output (JSON)	Supported	Supported	Supported
Image input	Not supported	Not supported	Not supported
Thinking mode	Supported	Supported	Supported
Price / 1M tokens
Input	$ 1.00>= 32K$ 1.20	$ 1.20	$ 0.90>= 32K$ 1.10
Output	$ 3.80>= 32K$ 4.40	$ 4.40	$ 3.70>= 32K$ 4.30
Cache read	$ 0.20>= 32K$ 0.30	$ 0.30	$ 0.18>= 32K$ 0.27
Cache write	$ 1.00>= 32K$ 1.20	$ 1.20	$ 0.90>= 32K$ 1.10
What it is good at
Best for	Long-horizon engineering Performance engineering	Million-token workflows Agentic software engineering	Persistent agent workflows
Published benchmarks
Terminal-Bench 2.1pass_rate	—	81%Terminus-2	—
SWE-Bench Proresolved_rate	58.4%(setup not fully disclosed)	62.1%(setup not fully disclosed)	—
SciCodescore	—	—	43.6%enabled(setup not fully disclosed)
AA-LCRscore	—	—	66.7%enabled(setup not fully disclosed)
KernelBench Level 3 Optimizationgeometric_mean_speedup	3.6×(setup not fully disclosed)	—	—
AA Intelligence Index v4.1.1score	—	—	39.1enabled(setup not fully disclosed)
GPQA Diamondaccuracy	—	91.2%(setup not fully disclosed)	—
Humanity's Last Examaccuracysetups differ by: reasoning, tools	—	40.5%	27.8%enabled(setup not fully disclosed)
Humanity's Last Exam (with tools)accuracy	—	54.7%	—
CritPtscore	—	—	0.3%enabled(setup not fully disclosed)
GPQA Diamondaccuracy	—	—	84.7%enabled(setup not fully disclosed)
MCP Atlas Publicscore	—	76.8%(setup not fully disclosed)	—
Tool Decathlonscore	—	48.2%(setup not fully disclosed)	—