Skip to content
ShilpBot
AI Models

Baidu ERNIE-Light-DT-1.0: Pricing, Specs and Limits

Baidu ERNIE-Light-DT-1.0: reasoning model based on open-source GLM models with Baidu Sparse Attention, 1M context, 1.13 in / 3.94 out per 1M tokens....

By ShilpBot Team · Published · 3 min read

ERNIE-Light-DT-1.0 is based on open-source GLM models and includes research from Baidu Search's Lightning large model. Its listing highlights Baidu Sparse Attention (BSA), a sparse attention mechanism developed by Baidu.

Sparse attention is a technique for handling long inputs more efficiently, so this model is pitched at long-context work. It accepts text only.

Prices and specifications checked against the public OpenRouter catalogue on 2026-11-16. See the live model page for current limits.

Baidu ERNIE-Light-DT-1.0 quick facts

DetailValue
DeveloperBaidu
Model IDbaidu/ernie-light-dt-1.0-intl
Added to the catalogue2026-11-10
TypeReasoning model based on open-source GLM models with Baidu Sparse Attention
ParametersNot listed
Context windowabout 1M tokens (1,024,000)
Maximum output131,072 tokens
Input typestext
Price1.13 in / 3.94 out per 1M tokens (cached input 0.11)

Feature support from the catalogue: Tool calling: listed; JSON / structured output: listed; Adjustable reasoning effort: listed.

What Baidu ERNIE-Light-DT-1.0 is good for

  • Long-text reasoning with a 1,024,000-token window.
  • Search-flavoured tasks, given its lineage in Baidu Search research.
  • Text-only pipelines where multimodal input is not needed.

Limits of Baidu ERNIE-Light-DT-1.0

  • No image, audio or video input.
  • It costs more per input token ($1.13) than several multimodal competitors.
  • Name suffixes like "DT" and "intl" refer to the provider's internal naming. Check the model page for what the international listing includes.

How Baidu ERNIE-Light-DT-1.0 compares

  • Baidu ERNIE 5.5 Flash Preview: Both have a 1,024,000-token context window; its output cap is 131,072 tokens against 921,600; it is a paid model, while Baidu ERNIE 5.5 Flash Preview has a free listing.
  • Z.ai GLM-5.3-Prime: It has a larger context window (1,024,000 against 1,000,000 tokens); both cap output at 131,072 tokens; it costs 1.13 input and 3.94 output per million tokens against 2.80 and 8.80.
  • Nex AGI Nex-N2.6: It has a larger context window (1,024,000 against 262,144 tokens); its output cap is 131,072 tokens against 235,929; it does not accept image input; it costs 1.13 input and 3.94 output per million tokens against 0.12 and 0.40.

Calling Baidu ERNIE-Light-DT-1.0 from code

Use the model ID baidu/ernie-light-dt-1.0-intl with any OpenAI-compatible client pointed at OpenRouter:

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "baidu/ernie-light-dt-1.0-intl", "messages": [{"role": "user", "content": "Explain what you can do in two sentences."}]}'

Prompts that show what Baidu ERNIE-Light-DT-1.0 can do

  1. Paste a long article collection and ask for the three strongest disagreements between sources.
  2. Ask for a step-by-step plan for a research project and check the order of steps.
  3. Compare it with ERNIE 5.5 Flash Preview (free) on the same task.

What Baidu ERNIE-Light-DT-1.0 costs

The catalogue lists 1.13 per million input tokens and 3.94 per million output tokens, with cached input at 0.11 per million. Sending 1 million input tokens and receiving 100,000 output tokens would cost about 1.52. A single coding-agent step with a 150,000-token prompt and a 10,000-token answer would cost about $0.21 before caching. If your client reuses the same prompt prefix, cached input at 0.11 per million tokens can cut the input side of that bill by about 90 percent. Prices are set by the provider, can change, and may differ between hosting providers, so check the live model page before you budget.

Keep reading

Compare every paid model by price in our AI model pricing roundup, or see the free models roundup. New to terms like mixture-of-experts? Start with how neural networks learn.

Frequently asked questions

Is there a free ERNIE model?

Yes. The free ERNIE-5.5-Flash-Preview is covered on our site and also offers about 1M tokens of context.

What are the cost and limits of Baidu ERNIE-Light-DT-1.0?

As of 2026-11-16 the catalogue lists 1.13 per million input tokens and 3.94 per million output tokens. About 1M tokens (1,024,000) context window, up to 131,072 output tokens, model ID baidu/ernie-light-dt-1.0-intl. Prices can change, so check the live model page.

#Baidu ERNIE-Light-DT-1.0#Baidu model#AI model pricing#AI API pricing#AI chat model#reasoning model#OpenRouter