ichq7-002

GMP KnowledgeIch Q7medium

Question

Under ICH Q7, what is the requirement for process validation of API manufacturing processes?

Choices

A. Process validation is only required for sterile APIs
B. Process validation should provide documented evidence that the process, operated within established parameters, can perform effectively and reproducibly to produce an API meeting specifications
C. Three consecutive validation batches are always mandatory regardless of the process type
D. Process validation may be replaced entirely by in-process testing

Cross-Model Comparison

Model	Score	Latency	Tokens In	Tokens Out
Qwen3.8-Flash	N/A	N/A	-	-
GLM-5.3	N/A	N/A	-	-
MiniMax M3	N/A	N/A	-	-
Qwen3.8 Max (2.4T-A95B)	N/A	N/A	-	-
Hy4 Preview	N/A	N/A	-	-
Muse Glimmer 30B	N/A	N/A	-	-
Llama 4 Scout	100.0%	1.6s	97	429
Llama 4 Maverick	100.0%	13.4s	98	414
Mistral Small 2603	100.0%	2.4s	102	373
Qwen3.5-397B-A17B	100.0%	33.3s	98	1,189
DeepSeek-V3.2	100.0%	21.9s	91	174
DeepSeek-R1	100.0%	21.3s	97	792
DeepSeek-R1-Distill-Qwen-32B	100.0%	27.2s	91	625
Mistral Large 3 675B	100.0%	5.9s	90	382
Gemini 3.1 Pro	100.0%	9.2s	88	871
Gemini 3 Flash	100.0%	2.6s	88	237
Gemini 3.1 Flash-Lite	100.0%	2.0s	88	243
Qwen3.6 27B	100.0%	12.9s	98	1,053
Qwen3.6 35B A3B	100.0%	5.1s	98	924
DeepSeek V4 Pro	100.0%	16.9s	91	928
DeepSeek-V3.2	100.0%	8.1s	91	153
Llama 4 Scout	100.0%	8.2s	155	326
DeepSeek V4 Flash	100.0%	6.9s	91	219
Gemma 4 26B A4B IT	100.0%	4.0s	101	277
Gemma 4 31B IT	100.0%	8.7s	101	208
MiniMax M2.7	100.0%	7.2s	128	355
Claude Opus 4.8	100.0%	9.4s	150	678
GLM-5.2	100.0%	5.2s	100	423
Llama 4 Maverick	100.0%	20.7s	155	365
Llama 3.3 70B Instruct	100.0%	7.5s	183	299
Qwen3.8 27B	100.0%	2.5s	202	109
Mistral Small 2603	100.0%	2.2s	165	287
Mistral Large 3 675B	100.0%	3.7s	153	313
Qwen3.5-35B-A3B	100.0%	5.9s	160	1,178
Qwen3.6 27B	100.0%	44.6s	160	1,972
Qwen3.6 35B A3B	100.0%	6.2s	160	1,235
Qwen3.5-397B-A17B	100.0%	27.9s	160	2,120
DeepSeek V4 Pro	100.0%	5.2s	151	248
DeepSeek-V3.2	100.0%	5.8s	157	139
DeepSeek V4 Flash	100.0%	3.9s	151	275
DeepSeek-R1	100.0%	27.5s	150	887
Gemma 4 26B A4B IT	100.0%	7.2s	163	311
Gemma 4 26B A4B IT	100.0%	9.4s	163	210
Gemma 4 31B IT	100.0%	14.5s	163	229
Hy4 Preview	100.0%	50.2s	168	4,200
MiniMax M2.7	100.0%	6.3s	188	314
GLM-5.2	100.0%	13.0s	160	707
Kimi K3	100.0%	16.0s	232	567
GLM-5.3-Flash	100.0%	33.1s	160	1,053
GPT-5.6 Luna	100.0%	2.1s	151	101
GPT-5.4 mini	100.0%	1.0s	93	67
GPT-5.4 nano	100.0%	1.9s	93	60
Claude Haiku 4.5	100.0%	3.5s	98	304
Claude Sonnet 4.6	100.0%	9.0s	98	392
Claude Opus 4.6	100.0%	11.1s	98	417
GPT-5.6 Terra	100.0%	2.4s	151	79
GPT-5.4	100.0%	2.6s	93	154
Muse Glimmer 30B	100.0%	6.0s	201	363
Qwen3.8 Max (2.4T-A95B)	100.0%	2.3s	202	192
Qwen3.8-Flash	100.0%	3.4s	211	283
MiniMax M3	100.0%	6.9s	310	221
GLM-5.3	100.0%	7.8s	160	1,135
Claude Opus 5	100.0%	12.1s	238	788
Claude Sonnet 5	100.0%	1.4s	238	9
Claude Haiku 4.5	100.0%	3.0s	162	289
GPT-5.6 Sol	100.0%	1.6s	151	45
Qwen3.5-35B-A3B	0.0%	8.7s	98	1,468
Llama 3.3 70B Instruct	0.0%	1800.5s	99	4

ichq7-002

Cross-Model Comparison

Tags