AI models
CogAgent
Training compute
6.7×10²² FLOP
Parameters
18B
Published
Dec 14, 2023
CogAgent is an AI model developed by Tsinghua University and Z.ai (Zhipu AI) (China), first published in December 2023. It works in the vision and language domain, on tasks such as instruction interpretation and visual question answering.
Training it took an estimated 6.7×10²² FLOP of compute (estimation method: operation counting). The model has 18,000,000,000 parameters.
Access: Open weights (restricted use). Its weights are openly available. It is built on top of CogVLM-17B. The reference paper has 731 citations. Epoch AI rates the confidence of this record as likely.
Full record
- Organization
- Tsinghua University, Z.ai (Zhipu AI)
- Country of organization
- China
- Domain
- Vision, Language
- Task
- Instruction interpretation, Visual question answering
- Training compute
- 6.7×10²² FLOP
- Compute estimation method
- Operation counting
- Parameters
- 18,000,000,000
- Numerical format
- BF16
- Model accessibility
- Open weights (restricted use)
- Open weights
- Yes
- Base model
- CogVLM-17B
- Citations
- 731
- Epoch confidence
- Likely
SourceEpoch AI, 'AI Models'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/ai-models. Licensed under CC BY 4.0.
← All ai models