annex1-001

GMP KnowledgeEu Annex1easy

Question

According to EU GMP Annex 1 (2022 revision), what is the maximum permitted total particle count for Grade A zones at rest for particles >= 0.5 um per m3?

Choices

A. 3,520
B. 35,200
C. 352,000
D. 3,520,000

Cross-Model Comparison

Model	Score	Latency	Tokens In	Tokens Out
GLM-5.3	N/A	N/A	-	-
Qwen3.8-Flash	N/A	N/A	-	-
Muse Glimmer 30B	N/A	N/A	-	-
Hy4 Preview	N/A	N/A	-	-
MiniMax M3	N/A	N/A	-	-
Qwen3.8 Max (2.4T-A95B)	N/A	N/A	-	-
Llama 4 Maverick	100.0%	16.2s	80	306
DeepSeek V4 Flash	100.0%	3.1s	73	416
Llama 3.3 70B Instruct	100.0%	6.6s	104	54
Mistral Small 2603	100.0%	1.3s	100	206
Qwen3.5-397B-A17B	100.0%	25.3s	95	1,547
DeepSeek-V3.2	100.0%	5.5s	79	66
DeepSeek-R1	100.0%	12.6s	79	439
DeepSeek-R1-Distill-Qwen-32B	100.0%	31.1s	87	741
Mistral Large 3 675B	100.0%	3.2s	88	229
Qwen3.5-35B-A3B	100.0%	12.1s	95	1,996
Gemini 3.1 Pro	100.0%	7.7s	84	843
Gemini 3 Flash	100.0%	1.6s	84	145
Gemini 3.1 Flash-Lite	100.0%	1.1s	85	79
Qwen3.6 27B	100.0%	34.3s	95	2,565
Qwen3.6 35B A3B	100.0%	6.7s	95	1,369
DeepSeek V4 Pro	100.0%	8.4s	73	470
Gemma 4 26B A4B IT	100.0%	6.1s	97	301
Gemma 4 31B IT	100.0%	6.4s	97	147
MiniMax M2.7	100.0%	5.7s	109	242
Llama 4 Maverick	100.0%	6.1s	137	184
Claude Opus 4.8	100.0%	5.6s	95	191
GLM-5.2	100.0%	8.1s	82	425
Llama 4 Scout	100.0%	6.3s	137	212
Llama 3.3 70B Instruct	100.0%	5.3s	140	219
Mistral Small 2603	100.0%	2.2s	163	243
Mistral Large 3 675B	100.0%	4.8s	151	179
Qwen3.5-35B-A3B	100.0%	8.7s	157	1,908
Kimi K3	100.0%	14.4s	214	470
Qwen3.6 27B	100.0%	58.5s	157	2,178
Qwen3.6 35B A3B	100.0%	59.0s	157	1,620
Qwen3.5-397B-A17B	100.0%	17.9s	157	1,834
DeepSeek V4 Pro	100.0%	6.3s	133	250
DeepSeek-V3.2	100.0%	6.1s	139	144
MiniMax M3	100.0%	669ms	304	75
DeepSeek V4 Flash	100.0%	2.9s	133	152
DeepSeek-R1	100.0%	110.6s	134	2,064
Gemma 4 26B A4B IT	100.0%	7.3s	159	394
Gemma 4 31B IT	100.0%	7.6s	159	287
MiniMax M2.7	100.0%	12.2s	169	1,177
Qwen3.8 Max (2.4T-A95B)	100.0%	3.6s	199	447
GLM-5.2	100.0%	5.7s	142	443
Muse Glimmer 30B	100.0%	19.0s	183	1,121
Qwen3.8 27B	100.0%	18.9s	195	1,069
GLM-5.3-Flash	100.0%	21.5s	142	833
Hy4 Preview	100.0%	86.5s	150	7,460
GPT-5.6 Luna	100.0%	1.9s	132	150
GPT-5.4 mini	100.0%	999ms	74	65
Claude Haiku 4.5	100.0%	2.4s	82	204
Claude Sonnet 4.6	100.0%	7.2s	82	364
Claude Opus 4.6	100.0%	10.2s	82	360
Llama 4 Scout	100.0%	1.1s	79	293
GPT-5.6 Terra	100.0%	2.8s	132	121
GPT-5.4	100.0%	1.8s	74	71
Qwen3.8-Flash	100.0%	5.9s	208	499
GLM-5.3	100.0%	8.4s	142	578
Claude Opus 5	100.0%	3.4s	183	111
Claude Sonnet 5	100.0%	2.4s	183	48
Claude Haiku 4.5	100.0%	3.0s	146	162
GPT-5.6 Sol	100.0%	2.7s	132	108
DeepSeek-V3.2	0.0%	7.3s	73	175
GPT-5.4 nano	0.0%	732ms	74	58

annex1-001

Cross-Model Comparison

Tags