Baidu sources
3 official pages from Baidu support 1 model record in the AI model benchmarks. Official sources come from the model providers themselves: they support each model record as a whole, and no benchmark score comes from them.
Each record names the date it was last verified, its dataset version and license where Baidu states one, what it covers and how the data uses it. Missing metadata is not filled or inferred. Every publisher is listed in the source catalog.
Baidu 3
- Introducing ERNIE 5.5 Flash Preview
Covers identity, benchmark · Verified 2026-11-11 · Not versioned · License not stated
Published 2026-11-11. The first post-trained model on ERNIE 5.5 Flash Base: 280B total parameters with 14B active and a 1M-token context, text only; the API is on Qianfan and Baidu says the weights will be open-sourced under Apache 2.0 soon. Its appendix table compares it with Hy4, DeepSeek-V4.1-Flash-1017, GLM-5.3-Flash, GLM-5.3, GPT-6.1 Sol and Claude Opus 5.5 at their highest reasoning setting; scores marked with an asterisk are Baidu runs.
- Qianfan model list (international)
Covers identity · Verified 2026-11-11 · Not versioned · License not stated
ERNIE 5.5 Flash, version ERNIE-5.5-Flash-Preview, model parameter ernie-5.5-flash-preview: 1M context length, 1M maximum input and 1M maximum output tokens, listed for text generation and for deep thinking (no thinking_budget).
- Qianfan model inference pricing (international)
Covers pricing · Verified 2026-11-11 · Not versioned · License not stated
Read 2026-11-11. Pay as you go, online inference: ERNIE-5.5-Flash-Preview at $0.28 input, $0.028 cache hit and $1.13 output per million tokens.