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-5.3-FlashGLM-5.3GLM-5.2 Add a modelChat with these 3

AttributeHide identical rows	GLM-5.3-Flashz-ai/glm-5.3-flash50% off	GLM-5.3z-ai/glm-5.3	GLM-5.2z-ai/glm-5.2
Specs
Context window	1M	1M	1M
Max output	128K	128K	128K
Input modalities	TextImageVideo	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	Supported	Not supported	Not supported
Thinking mode	Supported	Supported	Supported
Price / 1M tokens
Input	$ 0.15$ 0.075	$ 1.20	$ 1.20
Output	$ 0.50$ 0.25	$ 4.40	$ 4.40
Cache read	$ 0.030$ 0.015	$ 0.30	$ 0.30
Cache write	$ 0.15$ 0.075	$ 1.20	$ 1.20
What it is good at
Best for	Visual coding loop Long-horizon agentic engineering Office and financial documents	Long-horizon software engineering Security-focused code review	Million-token workflows Agentic software engineering
Published benchmarks
Terminal-Bench 2.1pass_ratesetups differ by: harness, max_output_tokens, temperature, top_p	84.3%Claude Code 2.1.207(setup not fully disclosed)	—	81%Terminus-2
Terminal-Bench 3.0pass_rate	—	28.3%effort max · Claude Code 2.1.207 · 3 runs(setup not fully disclosed)	—
Agents' Last Exam (CLI)scoresetups differ by: harness, max_output_tokens	26.3%effort max · Claude Code(setup not fully disclosed)	28.5%effort max(setup not fully disclosed)	—
AutomationBench v1.0.6score	48.8%(setup not fully disclosed)	—	—
DeepSWE 1.1resolved_rate	63.4%mini-swe-agent(setup not fully disclosed)	66.9%mini-swe-agent(setup not fully disclosed)	—
NL2Reposcore	56.3%(setup not fully disclosed)	—	—
SWE-Bench Proresolved_rate	—	—	62.1%(setup not fully disclosed)
CyberGymscore	—	84.5%effort max · Claude Code 2.1.207(setup not fully disclosed)	—
ExploitBenchscore	—	54.4%effort max · Claude Code 2.1.207 · 3 runs(setup not fully disclosed)	—
ExploitGym (6h)tasks_completed	—	130effort max · Claude Code 2.1.207(setup not fully disclosed)	—
OfficeQA Proscore	62.4%(setup not fully disclosed)	—	—
GDPval-AA v2elo	1773(setup not fully disclosed)	1769(setup not fully disclosed)	—
AA Intelligence Index v4.1.1score	57enabled(setup not fully disclosed)	—	—
GPQA Diamondaccuracy	—	—	91.2%(setup not fully disclosed)
Humanity's Last Examaccuracy	—	—	40.5%
Humanity's Last Exam (with tools)accuracysetups differ by: max_output_tokens, reasoning, temperature, tools, top_p	55.3%(setup not fully disclosed)	62.5%effort max(setup not fully disclosed)	54.7%
MCP Atlas Publicscore	—	—	76.8%(setup not fully disclosed)
Tool Decathlonscore	—	—	48.2%(setup not fully disclosed)
Toolathlon Verifiedpass_rate	78.4%3 runs(setup not fully disclosed)	—	—
MMVUaccuracy	80.5%(setup not fully disclosed)	—	—