Doubao-Seed 1.6 Flash

Doubao-Seed 1.6 Flash sits within ByteDance's Doubao model family as a lightweight variant engineered for speed and scale. Reseller documentation describes it as an extremely fast multimodal deep thinking model, with a reported time-per-output-token of roughly 10ms that places it firmly in latency-sensitive territory. Its multimodal design covers text alongside visual understanding for images and video, with text comprehension positioned above the prior-generation lite model and visual comprehension reported to match competing pro-tier offerings.

The model's design priorities make it a strong fit for high-throughput, real-time applications where response speed matters more than maximum reasoning depth. Practical deployment scenarios include high-concurrency APIs, lightweight assistants, coding agents that benefit from long-context recall, and any pipeline where multimodal inputs arrive as a steady stream. Because its strengths center on fast inference and broad input handling rather than the deepest analytical chains, it works best as a responsive front-line model that can hand off more complex reasoning tasks to heavier variants when needed.

Qiniudoubao-seed-1.6-flashseeddeprecated

Open provider page Provider docs

Quick Info

Provider: Qiniu
Model key: doubao-seed-1.6-flash
Release date: Aug 28, 2025
Last updated: Aug 28, 2025
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $0.021739
Output token cost: $0.217391

Limits

Output tokens: 32,000 tokens
Context window: 256,000 tokens

Transparent token rates

Compare Doubao-Seed 1.6 Flash pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Doubao-Seed 1.6 Flash

Qiniu

Coverage

The Baidu Baike entry on the Doubao Large Model states that it is an integrated speech understanding and generation large model developed by ByteDance and originally launched under the name "Lark" via Volcano Engine, with subsequent milestones including multimodal speech, 128K long context, and end-to-end voice dialogu The entry further describes ByteDance-side milestones such as the January 2025 launch of the Doubao Real-time Speech Large Model, the formation of the "Seed Edge" AGI team, the February release of the open-source VideoWorld model and the UltraMem sparse architecture, the March COMET communication optimization system, a

Qiniu

Coverage

A June 10, 2025 news flash reports that ByteDance's Volcano Engine officially released Doubao Large Model 1.6 at the Volcanic Engine Conference on June 11, with the lineup including three variants: Doubao-Seed-1.6-thinking, Doubao-Seed-1.6, and Doubao-Seed-1.6-flash. The thinking variant is highlighted as having enhanc The report also cites a Doubao large model score of 144 on a national new-style mathematics college entrance examination, ranking first nationally. The page is presented as breaking news translated by third-party software and carries a financial-news disclaimer rather than first-party ByteDance technical documentation,

Sulat also reports on AI and tech beyond the catalog.Read news.sulat.com