A Guth Labs publication ↗
Back to News

Baidu lists ERNIE 5.5 Flash Preview for API access with a million-token context

Baidu’s Qianfan documentation now lists ERNIE-5.5-Flash-Preview as an API model, with a one-million-token context window. The catalog, now dated November 11, confirms the service specification. Baidu model catalog

The catalog includes the preview in both text-generation and reasoning sections. Update checked November 15: OpenRouter now redirects the archived model link to a free preview listing, retaining a November 10 release date. The earlier draft’s paid OpenRouter rates are no longer presented as current. Baidu model catalog OpenRouter listing

Baidu’s own pricing table shows promotional rates alongside regular prices. At the published peak-hour promotional rate, input costs 1 yuan per million tokens and output costs 4 yuan; off-peak promotional rates are half those amounts. The original prices are twice the promotional rates. The displayed table does not establish how long that promotion will last. Baidu pricing

The API preview should be distinguished from ERNIE 5.5 Flash Base, the earlier pretraining checkpoint intended for further adaptation. An accessible model endpoint does not by itself establish that the preview’s downloadable weights or a matching license are publicly available. Baidu base-model description

For prospective users, the useful next comparison is workload-specific quality, latency and actual billing. A large context limit and a published token price provide deployment parameters, but do not establish reliable long-document reasoning or a competitive performance ranking.

Sources

Sources and citations

Revision history

Published .