PLPramaana

Z.ai

GLM 5.2

Release date: 6/13/2026

An open-weight bilingual reasoning model tuned for code and long-context analysis.

Pramaana IndexAccuracy: StandardLatency: StrongCost: Leading
Accuracy / Pramaana Index
64.79%± 1.67
Latency / Pramaana Index
348.43s
Cost / Test / Pramaana Index
$0.37
Context window
256k
Max output tokens
64k
Input modality
Text · Image
Hyperparameter settings

temperature: 0.2

reasoning: extended

fallbacks: disabled

Benchmark performance for GLM 5.2

BenchmarksAccuracyRanking
Pramaana Index64.79% ± 1.677 / 35
Pramaana Multimodal Index60.71% ± 0.489 / 24
Legal Research Fidelity59.19% ± 0.657 / 18
Financial Modeling Agent57.30% ± 0.826 / 22
Clinical Documentation Safety58.49% ± 0.998 / 19
MATH-50072.59% ± 1.1610 / 112
GPQA Diamond76.60% ± 1.339 / 121
MMLU-Pro74.09% ± 1.508 / 119
Adaptive Tutoring Dialogue60.39% ± 1.678 / 16
SWE-bench Verified56.10% ± 0.485 / 69
LiveCodeBench73.01% ± 0.656 / 127
Agentic Reliability48.91% ± 0.829 / 12
Strategic Games Arena57.00% ± 0.998 / 20
Benefits Navigator60.30% ± 1.168 / 16

License: Apache 2.0 · Evaluation settings, model endpoints, and benchmark versions are frozen with each published run.

Read our methodology