Baidu: ERNIE 4.5 21B A3B Thinking

Text input Text output
Author's Description

ERNIE-4.5-21B-A3B-Thinking is Baidu's upgraded lightweight MoE model, refined to boost reasoning depth and quality for top-tier performance in logical puzzles, math, science, coding, text generation, and expert-level academic benchmarks.

Key Specifications
Cost
$$$$
Context
131K
Parameters
21B
Released
Oct 09, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Stop Max Tokens Top P Reasoning Seed Frequency Penalty Presence Penalty Include Reasoning Temperature
Features

This model supports the following features:

Reasoning
Performance Summary

Baidu's ERNIE 4.5 21B A3B Thinking model demonstrates exceptional speed, consistently ranking among the fastest models across all benchmarks. It also offers highly competitive pricing, placing it among the most cost-effective options available. With an 87% success rate, the model exhibits strong reliability, indicating few technical issues during evaluation. Performance across benchmarks reveals a mixed but generally strong profile. The model excels in Reasoning, achieving 86.0% accuracy (75th percentile), aligning with its description of boosted reasoning depth. It also shows strong performance in General Knowledge (97.0% accuracy) and Ethics (98.0% accuracy). However, a significant weakness is observed in Instruction Following, where it scored 0.0% accuracy, suggesting a critical area for improvement. Coding performance is also low at 17.3% accuracy. Hallucinations are present at 80.0% accuracy, indicating room for improvement in acknowledging uncertainty. Email Classification and Mathematics show moderate performance at 96.0% and 68.0% accuracy, respectively. Overall, its key strengths lie in speed, cost-efficiency, and advanced reasoning capabilities, while instruction following and coding present notable weaknesses.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.056
Completion $0.224

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Novita
Novita | baidu/ernie-4.5-21b-a3b-thinking 131K $0.056 / 1M tokens $0.224 / 1M tokens
Novita
Novita | baidu/ernie-4.5-21b-a3b-thinking 131K $0.056 / 1M tokens $0.224 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by baidu