claude-sonnet-5 vs glm-5.2

Pricing, Performance & Features Comparison

claude-sonnet-5

Authoranthropic

Context Length1M

Reasoning

Providers1

ReleasedJun 2026

Knowledge Cutoff–

License–

Near-Opus quality on coding and agentic work at Sonnet cost; adaptive thinking is on by default and non-default sampling parameters are rejected.

Input$3

Output$15

Latency (p50)5.8s

Output Limit128K

Function Calling

JSON Mode

InputText, Image

OutputText

anthropic

in$3out$15cache$0.3write$3.8

Latency (24h)

Success Rate (24h)

glm-5.2

Authorzai

Context Length1M

Reasoning

Providers1

ReleasedJun 2026

Knowledge Cutoff–

License–

Z.ai’s flagship (744B MoE, MIT open weights) with a 1M-token context window and adjustable reasoning effort, built for long-horizon coding and agentic tasks.

Input$1.4

Output$4.4

Latency (p50)7s

Output Limit131K

Function Calling

JSON Mode

InputText

OutputText

zai

in$1.4out$4.4cache$0.26–