Author's Description
A sophisticated text-based Mixture-of-Experts (MoE) model featuring 21B total parameters with 3B activated per token, delivering exceptional multimodal understanding and generation through heterogeneous MoE structures and modality-isolated routing. Supporting an...
Key Specifications
Supported Parameters
This model supports the following parameters:
Performance Summary
Baidu's ERNIE 4.5 21B A3B, a sophisticated Mixture-of-Experts model with 21B total parameters and 3B activated per token, demonstrates a balanced performance profile. It offers competitive response times, ranking in the 58th percentile for speed, and consistently provides among the most competitive pricing, placing in the 82nd percentile. Notably, its reliability is exceptional, boasting a 98% success rate across benchmarks, indicating minimal technical failures. The model exhibits strengths in cost-efficiency across various tasks, consistently ranking in the top 20% for cost-effectiveness on most benchmarks. Its accuracy in "Hallucinations (Baseline)" is 92.0%, suggesting a good ability to acknowledge uncertainty. However, its accuracy in core cognitive areas like "General Knowledge" (94.3%), "Ethics" (94.5%), "Mathematics" (82.0%), "Email Classification" (93.5%), "Instruction Following" (51.2%), "Reasoning" (50.0%), and "Coding" (77.2%) generally falls in the lower percentiles (23rd-45th), indicating room for improvement in these specific domains compared to other models. Its extensive 120,000 context length and advanced post-training techniques are key features, but the benchmark results suggest that while it is highly reliable and cost-effective, its raw accuracy in complex reasoning and knowledge-based tasks is not its primary strength.
Model Pricing
Current Pricing
| Feature | Price (per 1M tokens) |
|---|---|
| Prompt | $0.07 |
| Completion | $0.28 |
Price History
Available Endpoints
| Provider | Endpoint Name | Context Length | Pricing (Input) | Pricing (Output) |
|---|---|---|---|---|
|
Novita
|
Novita | baidu/ernie-4.5-21b-a3b | 120K | $0.07 / 1M tokens | $0.28 / 1M tokens |
|
Novita
|
Novita | baidu/ernie-4.5-21b-a3b | 120K | $0.07 / 1M tokens | $0.28 / 1M tokens |
Benchmark Results
| Benchmark | Category | Reasoning | Strategy | Free | Executions | Accuracy | Cost | Duration |
|---|
Other Models by baidu
|
|
Released | Params | Context |
|
Speed | Ability | Cost |
|---|---|---|---|---|---|---|---|
| Baidu: ERNIE 4.5 21B A3B Thinking | Oct 09, 2025 | 21B | 131K |
Text input
Text output
|
★ | ★★ | $$$$ |
| Baidu: ERNIE 4.5 VL 28B A3B | Aug 12, 2025 | 28B | 30K |
Text input
Image input
Text output
|
★★★ | ★★★ | $$$ |
| Baidu: ERNIE 4.5 VL 424B A47B | Jun 30, 2025 | 424B | 123K |
Text input
Image input
Text output
|
★★ | ★★★ | $$$$ |
| Baidu: ERNIE 4.5 300B A47B | Jun 30, 2025 | 300B | 123K |
Text input
Text output
|
★ | ★★ | $$$ |