shBench

#	Модель	Счёт
1	Gemini 3 Flash	10
1	Gemini 3 Pro	10
2	Gemini 2.5 Pro Deep Think	9
2	Gemini 2.5 Pro Deep Think (@loxAIbot)	9
2	GPT-5 Pro	9
3	GPT-5.2	8.5
4	Claude 4.5 Opus	8
4	GPT-5 Thinkng (Extended)	8
5	ERNIE 5.0 Preview	7.5
5	GPT 5.1 Thinking	7.5
5	Grok 4	7.5
5	Grok 4.1 Thinking	7.5
6	Claude 4.1 Opus	7
6	Gemini 2.5 Pro (AI Studio)	7
6	Gemini Robotics-ER 1.5 Preview	7
6	Manus 1.5	7
6	o3	7
7	Claude 4.5 Sonnet	6.5
7	Gemini 2.5 Flash (AI Studio)	6.5
7	Grok 4 Fast	6.5
7	Grok 4.1	6.5
7	o3-pro (genspark.ai)	6.5
7	Qwen 3 Max Thinking	6.5
8	GLM-4.6	6
8	o4 Mini High	6
9	Kimi K2 Thinking	5.5
9	Mistral LeChat	5.5
9	o1-pro (yupp.ai)	5.5
9	Qwen 3 Max	5.5
10	Claude 4 Sonnet	5
10	Claude 4.5 Haiku	5
10	GLM-4.5	5
10	GPT OSS 120B	5
10	Grok 4 Heavy	5
10	Nano Banana	5
10	o4 Mini	5
11	GPT 4.1	4.5
11	GPT 5.1 Instant	4.5
12	DeepSeek 3.2 Thinking	4
13	GigaChat	3.5
13	GPT 4o	3.5
13	MiniMax M2	3.5
14	Gemma 3 27B	3
14	GPT OSS 20B	3
14	Kimi K2	3
14	Ling-1T	3
14	Ring-1T	3
15	ChatGPT 5 Instant	2
15	GPT 4.5	2
15	YandexGPT	2
16	GPT 3.5 Turbo	1