Ai2 SciArena Benchmark Leaderboard | Phaseo

Ai2 SciArena

Type: numerical

View benchmark source

Recorded Results

16

Average Score

1034

Score Range

945 - 1172

Leading Model (lowest score)

945 - Mistral Small 3.1

Scores Over Time

Individual benchmark scores plotted by date.

Models Using This Benchmark

Organisation	Model	Reported	Top Score	Info	Self Reported	Source
Mistral	Mistral Small 3.1	17 Mar 2025	945	-	No	Source
Anthropic	Claude 3.7 Sonnet	24 Feb 2025	975	-	No	Source
Qwen	QwQ 32B	-	996	-	No	Source
Qwen	Qwen 3 32B	-	1009	-	No	Source
DeepSeek	DeepSeek V3 (2024-12-26)	26 Dec 2024	1016	-	No	Source
DeepSeek	DeepSeek R1 (2025-01-20)	20 Jan 2025	1018	-	No	Source
SpaceXAI	Grok 3	18 Apr 2025	1024	-	No	Source
OpenAI	GPT 4.1 Mini	14 Apr 2025	1026	-	No	Source
Qwen	Qwen 3 235B A22B	-	1026	-	No	Source
OpenAI	GPT 4.1	14 Apr 2025	1037	-	No	Source
Anthropic	Claude Sonnet 4	21 May 2025	1045	-	No	Source
OpenAI	o4 Mini	16 Apr 2025	1054	-	No	Source
DeepSeek	DeepSeek R1 (2025-05-28)	28 May 2025	1062	-	No	Source
Google	Gemini 2.5 Pro Preview (2025-06-05)	05 Jun 2025	1063	-	No	Source
Anthropic	Claude Opus 4	21 May 2025	1080	-	No	Source
OpenAI	o3	16 Apr 2025	1172	-	No	Source