Website Home (ChinaZ.com) June 14 News: ByteDance's Volcano Engine officially announced that its one-stop large model service platform, Volcano Ark, will now be open to individual developers. These developers can now experience and apply over a dozen large models, including the Doubao Model Family, as well as other platform features.
Services provided by Volcano Ark include:
The entire Doubao model series offers individual users a free inference quota of 500,000 tokens.
The 4K and 32K versions of the Doubao General Model pro and lite offer a traffic quota of 10,000 RPM (requests per minute) and 800,000 TPM (tokens per minute).
After the free quota is used up, the price for model inference input is 0.0008 yuan per thousand tokens.
Pricing example:
Doubao General Model pro-32k version, post-paid mode, billed by tokens used, with an hourly billing cycle.
Process for individual developers:
Complete personal real-name authentication on the Volcano Engine official website.
Enter the Volcano Ark console through the Volcano Ark/Doubao Large Model official website.
View the various large models currently offered in the Model Plaza, and check the model name, description, provider, whether it can be experienced, and application scenarios through the model card.
Model experience and comparison:
Developers can experience and compare models by operating on the Volcano Ark console page or through API calls.
Model fine-tuning:
Individual developers can upload datasets and optimize the model through SFT (Supervised FineTuning) or continued pre-training.
The opening of the Volcano Ark platform provides individual developers with a comprehensive and user-friendly large model service platform, which helps to lower the development threshold and accelerate the innovation and application of AI technology. Through Volcano Ark, developers can select suitable models for experience and application according to their needs, and can also make personalized adjustments and optimizations to the models to meet specific development needs.