On April 25th, Volcano Engine, a subsidiary of ByteDance, announced the full launch of its newly released Doubao 1.5 Deep Thinking model on its Edge Large Model Gateway, offering users a generous 5 million tokens of free usage. This move has garnered significant attention within the AI community.
Doubao 1.5 Deep Thinking model is a high-performance AI model developed by Volcano Engine. It excels in reasoning tasks within specialized fields such as mathematics, coding, and science, achieving or nearing globally top-tier performance. Beyond its exceptional reasoning capabilities, it demonstrates strong generalization abilities in creative writing and other non-reasoning tasks, enabling it to handle a wider range of complex scenarios. This release of Doubao 1.5 includes two versions: a large language model, Doubao-1.5-thinking-pro, with enhanced reasoning capabilities, and a multi-modal version supporting visual reasoning.
Volcano Engine's Edge Large Model Gateway provides robust support for the Doubao 1.5 Deep Thinking model. Users can now access the entire Doubao model series on the Edge Large Model Gateway and enjoy 2 million tokens of free usage. Furthermore, users who utilize the newly added Doubao large model in their gateway access keys can apply to increase their free quota to 5 million tokens.
The Edge Large Model Gateway uses a single API interface, allowing users to access models and agents from multiple large model providers. Leveraging its globally distributed edge computing nodes, it enables proximity calls, thereby improving model access speed and providing end-users with a faster and more reliable AI service experience. Users can access the Doubao 1.5 Deep Thinking model in two ways: directly using the pre-installed model on the Edge Large Model Gateway, or by incorporating third-party model platform keys into the Edge Large Model Gateway and accessing the model via the gateway's access keys.
The Doubao 1.5 Deep Thinking model employs a Mixture-of-Experts (MoE) architecture, boasting a total of 200 billion parameters while activating only 20 billion. This results in significant advantages in training and inference costs. Based on efficient algorithms, it offers extremely high concurrency capacity while achieving ultra-low latency of 20 milliseconds. Furthermore, its combined deep thinking and visual understanding capabilities make the Doubao 1.5 Deep Thinking model exceptionally efficient in problem-solving.
By accessing the Doubao 1.5 Deep Thinking model through the Edge Large Model Gateway, users benefit from accelerated calls, cost control, automatic retries, and failover. These features not only enhance service quality but also reduce operational costs, ensuring service continuity and reliability.
Currently, the Volcano Engine Edge Large Model Gateway supports over 100 mainstream large models and agents. To promote the Doubao 1.5 Deep Thinking model, the Edge Large Model Gateway has launched a special offer. Users who create access keys on the gateway and link the latest Doubao model can apply to increase their free quota to 5 million tokens.
This initiative by Volcano Engine not only provides powerful tools for AI developers but also strongly supports the popularization and application of AI technology. Through the Edge Large Model Gateway, users can more easily utilize the Doubao 1.5 Deep Thinking model, driving the application and development of AI technology in a wider range of fields.
