Yi 34B is an open-weight 34-billion parameter model from 01.AI, released to the public and available for download, fine-tuning, and self-hosting. It was trained with strong bilingual coverage of Chinese and English, and on release it ranked among the highest-performing open models in its size class on standard language understanding, coding, and reasoning benchmarks.
The model's open licensing and self-hostable nature make it a practical choice for organizations that require control over their inference environment or wish to fine-tune on proprietary data. Yi 34B has spawned numerous community fine-tunes targeting chat, roleplay, and domain-specific tasks, reflecting broad adoption across the open-source AI community.
Key Features
34B open-weight model available for self-hosting and fine-tuning
Strong bilingual Chinese and English language capability
Competitive reasoning and instruction-following for its size class
Active community ecosystem with many fine-tuned variants
Suitable for quantized deployment on consumer-grade GPUs
200K context window variant available for long-document tasks
Ideal Use Cases
Self-hosted inference where data sovereignty or privacy is required
Fine-tuning on domain-specific Chinese or English corpora
Research and experimentation on large open-weight models
Bilingual content generation and translation workflows
Base model for community or enterprise fine-tune projects
Example Prompts for Yi 34B
Technical Specifications
| Provider | Yi |
| Category | Text |
| Modality | Text -> Text |
Frequently Asked Questions
Try Yi 34B now
Start using Yi 34B instantly — 100 free credits, no credit card required. Access 750+ AI models through one platform.