Tasks and inputs
Seed 2.1 Lite is listed for Multimodal, Text → Text. It accepts text, image, audio, video and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Nov 17, 2026
Seed 2.1 Lite is a lightweight Doubao Seed 2.1 model listed on Volcengine Ark with a 1M context, 256K output, audio and image input and tool calls.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Indicative only, from catalogue listings: ~$0.115 in / $0.388 out per 1M tokens, ~$0.023 cache read, ~$1.73 audio input; one listing gives CNY 0.8 per 1M input. Not confirmed on a Volcengine page.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Cost-sensitive Chinese-market workloads are the obvious target for a "Lite" tier, and that is how this model is listed: a smaller member of the Doubao Seed 2.1 family on Volcengine Ark, versioned doubao-seed-2-1-lite-260915.
One caveat comes first. ByteDance's own Seed 2.1 materials from June 2026 describe Pro and Turbo tiers only. We could not find a first-party announcement of a Lite variant, so everything below comes from catalogue listings and should be checked in the Ark console before you build on it.
| Field | Listed value |
|---|---|
| Model ID | doubao-seed-2-1-lite-260915 |
| Context window | 1,024,000 tokens |
| Max output | 256,000 tokens |
| Input | Text, image, audio, video |
| Output | Text |
| Reasoning | Yes |
| Tool calling | Yes; GUI tasks also listed |
| Open weights | No |
The version suffix implies a 15 September 2026 build. That date is inferred, not announced.
These figures are not confirmed on a Volcengine page. Confirm on your own account.
Before migrating anything, call the ID from your Ark project and confirm it resolves. Then run a small set of your own prompts through it and through the earlier Seed 2.0 Lite, measuring quality, latency and cost per task. If the ID is not available in your region, Seed 2.0 Mini remains the documented low-cost option.
Small, cheap tiers in a family like this tend to be used for routing, classification, extraction and first-pass summarisation, with the heavier tier handling the hard cases. If the listed limits are accurate, a 1M-token window on a budget model is unusual and worth exploiting for long-document triage. Validate that claim directly: feed a long, known document and ask questions whose answers sit near the end.
Other Seed models are collected on the ByteDance Seed provider page.
Compare the model on the work you need to complete.
Seed 2.1 Lite is listed for Multimodal, Text → Text. It accepts text, image, audio, video and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
Documented limits: 1.02M context and 256K maximum output. The listed route has starting rates of $0.115 input and $0.388 output per million tokens. Measure a completed task using the prompts, documents and tools your application will use. Include reasoning and retries in the result.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
baidu
ERNIE 5.5 Flash Preview is Baidu's first post-trained 280B-A14B text model, with a 1M window, aimed at game dev, 3D scripting and knowledge work.
Meta
Muse Ember is Meta's smallest, cheapest Muse model: a 1M-context multimodal worker for tool calls, coding subtasks, routing and extraction at volume.
iflytek
Spark X2.6 is iFLYTEK's 293B mixture-of-experts language model family with integrated device and cloud variants, unveiled in October 2026.
Meta
Muse Spark Open is Meta's open-weight 816B mixture-of-experts model (42B active) with 1M context, built for coding and agents you can self-host.