llm-benchmarks
githubdrose.io

gmicloud Provider Benchmarks


Measured over the last 30 days

Models Tracked
23
Avg Tokens / Second
13.20
Avg Time to First Token (ms)
12324.57
Last Updated
Sep 2, 2026

All gmicloud models

23 tracked · fastest first
ModelAvg Toks/SecMinMaxAvg TTF (ms)
GLM 525.6025.6025.601900.00
GLM 5.220.8017.5025.602260.00
GLM 5.120.4019.6021.302260.00
GLM 5.118.900.4332.106790.00
MiniMax M318.8015.3023.303200.00
GLM 5.217.302.3230.803700.00
Qwen3 235B A22B Instruct 250716.4016.4016.402090.00
DeepSeek V3.216.304.0123.403320.00
DeepSeek V3.215.804.5521.104020.00
MiMo-V2.5-Pro14.9014.9014.903090.00
DeepSeek V3 032414.1014.1014.102730.00
GLM 512.505.1420.905700.00
DeepSeek V4 Flash 073112.104.5619.706120.00
DeepSeek V4 Flash 042310.006.5414.206490.00
DeepSeek V4 Flash 07319.003.9115.208670.00
MiMo-V2.58.825.3111.106890.00
MiMo-V2.5-Pro7.897.897.895850.00
MiniMax M2.75.722.849.1513100.00
MiniMax M2.75.232.079.0314600.00
Hy33.663.244.2917300.00
DeepSeek V4 Pro 08132.762.762.7622000.00
Kimi K2.61.651.651.6536300.00
Kimi K2.7 Code1.291.291.2947700.00

Frequently Asked Questions

Which gmicloud model is fastest?
Based on recent tests, GLM 5 shows the highest average throughput among tracked gmicloud models.
How many recent measurements feed this dashboard?
This provider summary aggregates 3963 individual prompts measured across 3530 monitoring runs over the past month.