AI models
Spec-Drafter
Training compute
1.2×10²⁰ FLOP
Parameters
500M
Published
Oct 30, 2023
Spec-Drafter is an AI model developed by Peking University and Microsoft Research Asia (China), first published in October 2023. It works in the language domain, on tasks such as translation.
Training it took an estimated 1.2×10²⁰ FLOP of compute (estimation method: operation counting). The model has 500,000,000 parameters. It was trained on roughly 39.3B datapoints. Training ran on 8 NVIDIA V100.
Access: Open weights (non-commercial). Its weights are openly available. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Peking University, Microsoft Research Asia
- Country of organization
- China
- Domain
- Language
- Task
- Translation
- Training compute
- 1.2×10²⁰ FLOP
- Compute estimation method
- Operation counting
- Parameters
- 500,000,000
- Dataset size
- 39.3B
- Training hardware
- NVIDIA V100
- Chips used
- 8
- Training power draw
- 4.8 kW
- Model accessibility
- Open weights (non-commercial)
- Open weights
- Yes
- Epoch confidence
- Confident
More from Peking University,Microsoft Research Asia
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