TPSTokens Per Second

GLM-4.7

Zhipu AI🇨🇳 China
active

~400B open-weight reasoning model for math, multi-file engineering, and agentic tasks. 200K context.

Compare with other models →
Context window200K tokens
Input / 1M tokens$0.38
Output / 1M tokens$1.98

Version History

glm-4-7-launchmajor

GLM-4.7 launches.

Benchmark Scores

Full leaderboard →
95.7%
AIME 2025
1442.0 elo
Arena Elo
85.9%
GPQA
90.3%
Hallucination Rate
89.4%
LiveCodeBench
85.6%
MMLU-Pro
114.0 tokens_per_sec
Speed (tok/s)
73.8%
SWE-bench Verified