Scientific Reasoning Leaderboard

Top 20 models ranked by gpqa diamond

Rank	Model	Creator	GPQA Diamond	Price/1M	Speed
🥇	Gemini 3.1 Pro Preview Google	Google	0.9	$4.50	115 tok/s
🥈	GPT-5.4 (xhigh) OpenAI	OpenAI	0.9	$5.63	76 tok/s
🥉	GPT-5.3 Codex (xhigh) OpenAI	OpenAI	0.9	$4.81	72 tok/s
4	Gemini 3 Pro Preview (high) Google	Google	0.9	$4.50	112 tok/s
5	GPT-5.2 (xhigh) OpenAI	OpenAI	0.9	$4.81	64 tok/s
6	GPT-5.2 Codex (xhigh) OpenAI	OpenAI	0.9	$4.81	89 tok/s
7	Gemini 3 Flash Preview (Reasoning) Google	Google	0.9	$1.13	190 tok/s
8	Claude Opus 4.6 (Adaptive Reasoning, Max Effort) Anthropic	Anthropic	0.9	$10.00	55 tok/s
9	Qwen3.5 397B A17B (Reasoning) Alibaba	Alibaba	0.9	$1.35	57 tok/s
10	Gemini 3 Pro Preview (low) Google	Google	0.9	$4.50	114 tok/s
11	Grok 4.20 Beta 0309 (Reasoning) xAI	xAI	0.9	$3.00	198 tok/s
12	Kimi K2.5 (Reasoning) Kimi	Kimi	0.9	$1.20	35 tok/s
13	Grok 4 xAI	xAI	0.9	$6.00	44 tok/s
14	Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) Anthropic	Anthropic	0.9	$6.00	68 tok/s
15	GPT-5.4 mini (xhigh) OpenAI	OpenAI	0.9	$1.69	249 tok/s
16	MiniMax-M2.7 MiniMax	MiniMax	0.9	$0.53	43 tok/s
17	GPT-5.1 (high) OpenAI	OpenAI	0.9	$3.44	89 tok/s
18	DeepSeek V3.2 Speciale DeepSeek	DeepSeek	0.9	$0.00	—
19	MiMo-V2-Pro Xiaomi	Xiaomi	0.9	$0.00	—
20	Claude Opus 4.5 (Reasoning) Anthropic	Anthropic	0.9	$10.00	59 tok/s