Self-Reference Count
Self Reference
Pass rate
0%
Survived 6 out of 15 breakers
ERNIE-4.5-300B-A47B is a 300B parameter Mixture-of-Experts (MoE) language model developed by Baidu as part of the ERNIE 4.5 series. It activates 47B parameters per token and supports text generation in both English and Chinese. Optimized for high-throughput inference and efficient scaling, it uses a heterogeneous MoE structure with advanced routing and quantization strategies, including FP8 and 2-bit formats. This version is fine-tuned for language-only tasks and supports reasoning, tool parameters, and extended context lengths up to 131k tokens. Suitable for general-purpose LLM applications with high reasoning and throughput demands.
123,000 tokens
$0.28 /1M tokens
$1.10 /1M tokens
12,000
| Test | Category | Latest Result | Success Rate | |
|---|---|---|---|---|
| Self-Reference Count | Self Reference | 0% | ||
| Contradictory Premises | Logic Reasoning | 0% | ||
| Car Wash Dilemma | Logic Reasoning | 0% | ||
| The Missing A | Pattern Matching | 0% | ||
| The Compartment Trick | Logic Reasoning | 0% | ||
| 10-Step Instructions | Instruction Following | 11% | ||
| Broken Mug | Lateral Thinking | 25% | ||
| Horse Race Logic | Logic Reasoning | 25% | ||
| Coin Flip Paradox | Logic Reasoning | 50% | ||
| Silence Protocol | Instruction Following | 67% | ||
| Alice's Brother Problem | Logic Reasoning | 89% | ||
| Strawberry Problem | Character Counting | 100% | ||
| Reverse Word Test | Character Manipulation | 100% | ||
| Bullshit Detector | Epistemic Humility | 100% | ||
| Sycophancy Trap | Logic Reasoning | 100% |