← Back to Test Cases
According to 21 CFR 211.192, what is required when any unexplained discrepancy or failure of a batch to meet specifications is found?
Knowledge QAGeneral Gmpeasy
Cross-Model Comparison
| Model | Score | Latency | Tokens In | Tokens Out |
|---|---|---|---|---|
| Gemini 3.1 Flash-Lite | 100.0% | 1.4s | 102 | 143 |
| GPT-5.4 mini | 100.0% | 736ms | 104 | 30 |
| GPT-5.4 nano | 100.0% | 1.4s | 104 | 69 |
| Claude Haiku 4.5 | 100.0% | 2.4s | 115 | 273 |
| DeepSeek-V3.2 | 100.0% | 24.2s | 101 | 109 |
| Claude Sonnet 4.6 | 100.0% | 8.1s | 115 | 322 |
| Llama 4 Scout | 100.0% | 5.1s | 108 | 401 |
| Claude Opus 4.6 | 100.0% | 9.2s | 115 | 332 |
| Llama 4 Maverick | 100.0% | 6.4s | 109 | 406 |
| Llama 3.3 70B Instruct | 100.0% | 1.9s | 110 | 178 |
| Mistral Small 2603 | 100.0% | 1.7s | 119 | 279 |
| Qwen3.5-397B-A17B | 100.0% | 12.7s | 115 | 844 |
| DeepSeek-R1 | 100.0% | 12.1s | 107 | 406 |
| DeepSeek-R1-Distill-Qwen-32B | 100.0% | 52.1s | 107 | 615 |
| Mistral Large 3 675B | 100.0% | 3.4s | 107 | 199 |
| GPT-5.4 | 100.0% | 1.6s | 104 | 67 |
| Gemini 3 Flash | 100.0% | 2.0s | 102 | 138 |
| Gemini 3.1 Pro | 100.0% | 6.5s | 102 | 607 |
| Qwen3.5-35B-A3B | 0.0% | 6.0s | 115 | 1,039 |
Tags
deviationsinvestigationbatch_disposition