Hy4 flash Base
Pretrained checkpoint for continued training and adaptation. The Flash architecture uses a sparse Engram memory module.
- Capability
- —No ECI score yet
- Input
- —No listed price
- Output
- —No listed price
- Context
- —tokens
What it costs
No per-token price is published for this model yet.
No per-token price is listed for this model on models.dev.
Independent benchmarks
Epoch AI has not published a Capabilities Index score for Hy4 flash Base yet.
The details
- Lab
- Tencent
- Task
- Multimodal pretraining
- Published capabilities
- Pretrained checkpoint for continued training and adaptation. The Flash architecture uses a sparse Engram memory module.
- Model size
- Approximately 220B total / 12B active per token
- Access
- Downloadable under the Tencent Hy License
- Sources
- Model weights
- Released
- Nov 18, 2026
- Context window
- — tokens
- Inputs
- Text, Images
- Output
- Text
- Reasoning
- Unconfirmed
- Tool calling
- Unconfirmed
- Structured output
- Unconfirmed
More from Tencent
About Hy4 flash Base
How much does Hy4 flash Base cost?
Hy4 flash Base has no published per-token API price on models.dev yet.
How good is Hy4 flash Base?
Epoch AI has not published a Capabilities Index score for Hy4 flash Base yet.
Is Hy4 flash Base open source?
Yes. Tencent publishes the weights under the Tencent Hy License licence, so it can be downloaded and self-hosted.
What inputs does Hy4 flash Base support?
Hy4 flash Base accepts text and images. Pretrained checkpoint for continued training and adaptation. The Flash architecture uses a sparse Engram memory module.
When was Hy4 flash Base released?
Tencent released Hy4 flash Base on Nov 18, 2026.
What are the best alternatives to Hy4 flash Base?
The closest current models from other labs on capability, price and release date are Aion 1.0 Instruct 2.3B (Microsoft), Aurora2-400M-A20M (AuroraAI Research), DaisyRose-27B (RadixArk) and Olala Mini 7A1B (OVHcloud).