ERNIE X1.1
ernie-x1.1
ProprietaryMultimodal
About
ERNIE X1.1 is Baidu's second-generation deep-thinking reasoning model, unveiled at Wave Summit on September 9, 2025. Built on ERNIE 4.5 multimodal foundations with iterative hybrid reinforcement learning and self-distillation. Delivers significant gains over ERNIE X1: factual accuracy up 34.8%, instruction following up 12.5%, agentic capabilities up 9.6%. Surpasses DeepSeek R1-0528 and performs on par with GPT-5 and Gemini 2.5 Pro on key benchmarks. Qianfan lists both ernie-x1.1 and ernie-x1.1-preview endpoints with 64K context and up to 65,536 output tokens.
ERNIE X1.1 has a 64K-token context window.
ERNIE X1.1 input tokens at $0.15/1M, output at $0.59/1M.
Capabilities
VisionMultimodalReasoningFunction CallingTool UseStructured OutputsCode ExecutionPrompt CachingBatch APIAudioFine-tuning
Providers(1)
| Provider | Input (per 1M) | Output (per 1M) | Type | |
|---|---|---|---|---|
| Baidu Qianfan | $0.15 | $0.59 | Serverless |