ERNIE-4.5-300B-A47B is a cutting-edge 300B parameter Mixture-of-Experts (MoE) language model developed by Baidu, part of their advanced ERNIE 4.5 series. This model is engineered for high performance, activating 47B parameters per token to deliver efficient and powerful text generation. It excels in both English and Chinese language tasks, making it a versatile tool for a global audience. Optimized for high-throughput inference and efficient scaling, ERNIE-4.5-300B-A47B leverages a heterogeneous MoE structure with advanced routing and quantization strategies, including FP8 and 2-bit formats. It supports reasoning, tool parameters, and an extended context length of up to 131k tokens, with a maximum output of 4k tokens. Pricing is competitive at $0.28 per 1M input tokens and $1.10 per 1M output tokens. This model is fine-tuned specifically for language-only tasks, making it suitable for general-purpose LLM applications demanding high reasoning capabilities and throughput. Access this STARTER tier model on Multi AI.
✅ Best For
🚀 Capabilities
❌ Limitations
Specifications
| Provider | baidu |
| Context Window | 123,000 tokens |
| Max Output | 4,096 tokens |
| Minimum Plan | Balance |
Pricing
| Input Price | $0.2800 / 1M tokens |
| Output Price | $1.1000 / 1M tokens |
💡 With PRO subscription, cost is reduced by 20%