Reasoning · Benchmark profile

CritPt

Research-level physics reasoning benchmark with composite challenges designed to stress frontier scientific reasoning.

Ranking weightedCurrent / rollingActive95 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on CritPt

Leaderboard · 95 models

GPT-5.6 Sol leads at 32.3%, followed by GPT-5.5 Pro (30.6%) and GPT-5.6 Terra (30%).

Ranking · 1.5%ReasoningCurrent

#1Closed

GPT-5.6 Sol

OpenAI

gpt-5.6-sol

32.3%

Overall #4 · Context 1.05M tokens

#2Closed

GPT-5.5 Pro

OpenAI

gpt.5.5.pro

30.6%

Overall #26 · Context Unavailable

#3Closed

GPT-5.6 Terra

OpenAI

gpt-5.6-terra

30%

Overall #9 · Context 1.05M tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

OpenAIAnthropicMoonshot AIZ.AIGooglexAIMetaAlibaba CloudDeepSeek

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
GPT-5.6 Sol	#1	OpenAI	32.3%
GPT-5.5 Pro	#2	OpenAI	30.6%
GPT-5.6 Terra	#3	OpenAI	30%
Claude Fable 5	#4	Anthropic	28.6%
GPT-5.5	#5	OpenAI	27.1%
GPT-5.4	#6	OpenAI	23.4%
Kimi K3	#7	Moonshot AI	23.4%
Claude Opus 4.8	#8	Anthropic	20.9%
GLM-5.2	#9	Z.AI	20.9%
GPT-5.6 Luna	#10	OpenAI	20.6%
Gemini 3.1 Pro Preview	#11	Google	17.7%
Claude Sonnet 5	#12	Anthropic	16.9%
GPT-5.3-Codex	#13	OpenAI	16.9%
Grok 4.5	#14	xAI	15.4%
Muse Spark 1.1	#15	Meta	15.1%
Qwen3.7-Max	#16	Alibaba Cloud	13.4%
Gemini 3.5 Flash	#17	Google	13.1%
DeepSeek V4 Pro	#18	DeepSeek	12.9%
Claude Opus 4.6	#19	Anthropic	12.6%
Claude Opus 4.7	#20	Anthropic	12%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
GPT-5.6 Sol	#1	OpenAI	32.3%
GPT-5.5 Pro	#2	OpenAI	30.6%
GPT-5.6 Terra	#3	OpenAI	30%
Claude Fable 5	#4	Anthropic	28.6%
GPT-5.5	#5	OpenAI	27.1%
GPT-5.4	#6	OpenAI	23.4%
Kimi K3	#7	Moonshot AI	23.4%
Claude Opus 4.8	#8	Anthropic	20.9%
GLM-5.2	#9	Z.AI	20.9%
GPT-5.6 Luna	#10	OpenAI	20.6%
Gemini 3.1 Pro Preview	#11	Google	17.7%
Claude Sonnet 5	#12	Anthropic	16.9%
GPT-5.3-Codex	#13	OpenAI	16.9%
Grok 4.5	#14	xAI	15.4%
Muse Spark 1.1	#15	Meta	15.1%
Qwen3.7-Max	#16	Alibaba Cloud	13.4%
Gemini 3.5 Flash	#17	Google	13.1%
DeepSeek V4 Pro	#18	DeepSeek	12.9%
Claude Opus 4.6	#19	Anthropic	12.6%
Claude Opus 4.7	#20	Anthropic	12%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
OpenAI	7	25.8%
Moonshot AI	1	23.4%
Z.AI	1	20.9%
Anthropic	5	18.2%
Google	2	15.4%
xAI	1	15.4%
Meta	1	15.1%
Alibaba Cloud	1	13.4%
DeepSeek	1	12.9%

Benchmark score table (95 models)

One best score per model · higher is better

CritPt leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	32.3%
#2	GPT-5.5 Pro gpt.5.5.pro	OpenAI	closed	30.6%
#3	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	30%
#4	Claude Fable 5 claude-fable-5	Anthropic	closed	28.6%
#5	GPT-5.5 gpt-5.5	OpenAI	closed	27.1%
#6	GPT-5.4 gpt-5.4	OpenAI	closed	23.4%
#7	Kimi K3 kimi-k3	Moonshot AI	closed	23.4%
#8	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	20.9%
#9	GLM-5.2 glm-5-2	Z.AI	open	20.9%
#10	GPT-5.6 Luna gpt-5.6-luna	OpenAI	closed	20.6%
#11	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	17.7%
#12	Claude Sonnet 5 claude-sonnet-5	Anthropic	closed	16.9%
#13	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	16.9%
#14	Grok 4.5 grok-4.5	xAI	closed	15.4%
#15	Muse Spark 1.1 muse-spark-1-1	Meta	closed	15.1%
#16	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	13.4%
#17	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	13.1%
#18	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	12.9%
#19	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	12.6%
#20	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	12%
#21	GPT-5.2 gpt-5.2	OpenAI	closed	11.6%
#22	Muse Spark muse-spark	Meta	closed	11.3%
#23	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	10%
#24	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	10%
#25	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	9.3%
#26	Gemini 3 Pro gemini-3-pro	Google	closed	9.1%
#27	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	9.1%
#28	GPT-5.2 Codex gpt-5.2-codex	OpenAI	closed	8.7%
#29	Gemini 3 Flash gemini-3-flash	Google	closed	8.6%
#30	Grok 4.3 grok-4.3	xAI	closed	8%
#31	Kimi K2.6 kimi-k2.6	Moonshot AI	open	8%
#32	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	7.1%
#33	Grok 4.20 grok-4.20	xAI	closed	6.6%
#34	GPT-5 gpt-5	OpenAI	closed	5.7%
#35	GPT-5 Codex gpt-5-codex	OpenAI	closed	5.1%
#36	GPT-5.1 gpt.5.1	OpenAI	closed	4.9%
#37	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	4.6%
#38	GLM-5.1 glm-5-1	Z.AI	open	4.6%
#39	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	4.3%
#40	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	4%
#41	MiMo-V2.5 mimo-v2.5	Xiaomi	open	3.7%
#42	MiniMax M3 MiniMax-M3	MiniMax	—	3.7%
#43	Qwen3.6 Max qwen3.6-max	Alibaba Cloud	closed	3.7%
#44	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	3.1%
#45	Kimi K2.5 kimi-k2.5	Moonshot AI	—	3.1%
#46	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	3.1%
#47	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	3.1%
#48	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	2.9%
#49	Grok 4 Fast grok-4-fast	xAI	closed	2.9%
#50	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	2.9%
#51	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	2.6%
#52	Kimi K2 Thinking kimi-k2-thinking	Moonshot AI	open	2.6%
#53	GLM-5 glm-5	Z.AI	open	2%
#54	Grok 4 grok-4	xAI	closed	2%
#55	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	1.7%
#56	Qwen3 Max qwen3-max	Alibaba Cloud	closed	1.7%
#57	GLM-4.7 glm-4.7	Z.AI	open	1.7%
#58	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	1.7%
#59	Gemma 4 31B gemma-4-31b	Google	open	1.4%
#60	GPT-5 mini gpt-5-mini	OpenAI	closed	1.4%
#61	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	1.1%
#62	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	1.1%
#63	GLM-4.6 glm-4.6	Z.AI	open	1.1%
#64	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	1.1%
#65	MiniMax M2.5 minimax-m2.5	MiniMax	open	1.1%
#66	o3 o3	OpenAI	closed	1.1%
#67	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	1.1%
#68	Claude Sonnet 3.7 claude-3-7-sonnet	Anthropic	closed	0.9%
#69	MiniMax M2 minimax-m2	MiniMax	open	0.9%
#70	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	0.9%
#71	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	0.9%
#72	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	0.6%
#73	Grok 3 mini grok-3-mini	xAI	closed	0.6%
#74	MiniMax M2.7 minimax-m2-7	MiniMax	open	0.6%
#75	o4-mini o4-mini	OpenAI	closed	0.6%
#76	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	0.6%
#77	Qwen3.5 Omni Plus qwen3.5-omni-plus	Alibaba Cloud	closed	0.6%
#78	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	0.3%
#79	Command A+ command-a-plus	Cohere	open	0.3%
#80	Gemini 2.5 Flash gemini-2.5-flash	Google	closed	0.3%
#81	GLM-5-Turbo glm-5-turbo	Z.AI	closed	0.3%
#82	Ling 2.6 1T ling-2.6-1t	InclusionAI	open	0.3%
#83	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	0.3%
#84	MiniMax M2.1 minimax-m2.1	MiniMax	open	0.3%
#85	Mistral Small 4 mistral-small-2603	Mistral AI	open	0.3%
#86	o1 o1	OpenAI	closed	0.3%
#87	Claude Haiku 4.5 claude-haiku-4-5-20251001	Anthropic	closed	0%
#88	Claude Opus 4.1 claude-opus-4-1	Anthropic	closed	0%
#89	Gemma 4 12B gemma-4-12b	Google	open	0%
#90	Gemma 4 26B gemma-4-26b-a4b	Google	open	0%
#91	Grok Code Fast 1 grok-code-fast-1	xAI	closed	0%
#92	Kimi K2 0905 kimi-k2-0905	Moonshot AI	open	0%
#93	Mistral Large 3 mistral-large-2512	Mistral AI	open	0%
#94	Mistral Medium 3.5 mistral-medium-3-5	Mistral AI	open	0%
#95	Nova 2 Pro nova-2-pro-preview	Amazon	closed	0%

The top of this snapshot is led by GPT-5.6 Sol at 32.3%; third place is 2.3 points behind. The top-10 spread is 11.7 points.

Compare top models

GPT-5.6 Sol vs GPT-5.5 Pro GPT-5.5 Pro vs GPT-5.6 Terra GPT-5.6 Terra vs Claude Fable 5 Claude Fable 5 vs GPT-5.5

About CritPt

Definition and scoring

Organisation: CritPt authors
Category: Reasoning
Version: Not evaluated
Direction: higher is better
Ranking use: Weighted 1.5%
Contamination risk: Low

Research-level physics reasoning benchmark with composite challenges designed to stress frontier scientific reasoning. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.

Open benchmark source ↗

FAQ

Quick answers

What does CritPt measure?

Research-level physics reasoning benchmark with composite challenges designed to stress frontier scientific reasoning.

Which model scores highest on CritPt?

GPT-5.6 Sol by OpenAI currently leads with 32.3%.

How many models are evaluated on CritPt?

95 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

Yes. This family is ranking-weighted at 1.5% of its capability category. GPT-5.6 Sol is overall #4.

GPQA Diamond

Rolling · 96 results · Weighted 2.5%

Humanity's Last Exam

Rolling · 105 results · Weighted 2.5%

ARC-AGI-2

2 · 10 results · Weighted 2.5%

LiveBench

Rolling · 17 results · Weighted 2.5%

ARC-AGI-3

3 · 3 results · Not Weighted

BenchLM Reasoning prior

bench-align-v5.1 · 18 results · Weighted 1.1%

Reasoning · Benchmark profile

CritPt

Research-level physics reasoning benchmark with composite challenges designed to stress frontier scientific reasoning.

Ranking weightedCurrent / rollingActive95 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on CritPt

Leaderboard · 95 models

GPT-5.6 Sol leads at 32.3%, followed by GPT-5.5 Pro (30.6%) and GPT-5.6 Terra (30%).

Ranking · 1.5%ReasoningCurrent

#1Closed

GPT-5.6 Sol

OpenAI

gpt-5.6-sol

32.3%

Overall #4 · Context 1.05M tokens

#2Closed

GPT-5.5 Pro

OpenAI

gpt.5.5.pro

30.6%

Overall #26 · Context Unavailable

#3Closed

GPT-5.6 Terra

OpenAI

gpt-5.6-terra

30%

Overall #9 · Context 1.05M tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

OpenAIAnthropicMoonshot AIZ.AIGooglexAIMetaAlibaba CloudDeepSeek

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
GPT-5.6 Sol	#1	OpenAI	32.3%
GPT-5.5 Pro	#2	OpenAI	30.6%
GPT-5.6 Terra	#3	OpenAI	30%
Claude Fable 5	#4	Anthropic	28.6%
GPT-5.5	#5	OpenAI	27.1%
GPT-5.4	#6	OpenAI	23.4%
Kimi K3	#7	Moonshot AI	23.4%
Claude Opus 4.8	#8	Anthropic	20.9%
GLM-5.2	#9	Z.AI	20.9%
GPT-5.6 Luna	#10	OpenAI	20.6%
Gemini 3.1 Pro Preview	#11	Google	17.7%
Claude Sonnet 5	#12	Anthropic	16.9%
GPT-5.3-Codex	#13	OpenAI	16.9%
Grok 4.5	#14	xAI	15.4%
Muse Spark 1.1	#15	Meta	15.1%
Qwen3.7-Max	#16	Alibaba Cloud	13.4%
Gemini 3.5 Flash	#17	Google	13.1%
DeepSeek V4 Pro	#18	DeepSeek	12.9%
Claude Opus 4.6	#19	Anthropic	12.6%
Claude Opus 4.7	#20	Anthropic	12%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
GPT-5.6 Sol	#1	OpenAI	32.3%
GPT-5.5 Pro	#2	OpenAI	30.6%
GPT-5.6 Terra	#3	OpenAI	30%
Claude Fable 5	#4	Anthropic	28.6%
GPT-5.5	#5	OpenAI	27.1%
GPT-5.4	#6	OpenAI	23.4%
Kimi K3	#7	Moonshot AI	23.4%
Claude Opus 4.8	#8	Anthropic	20.9%
GLM-5.2	#9	Z.AI	20.9%
GPT-5.6 Luna	#10	OpenAI	20.6%
Gemini 3.1 Pro Preview	#11	Google	17.7%
Claude Sonnet 5	#12	Anthropic	16.9%
GPT-5.3-Codex	#13	OpenAI	16.9%
Grok 4.5	#14	xAI	15.4%
Muse Spark 1.1	#15	Meta	15.1%
Qwen3.7-Max	#16	Alibaba Cloud	13.4%
Gemini 3.5 Flash	#17	Google	13.1%
DeepSeek V4 Pro	#18	DeepSeek	12.9%
Claude Opus 4.6	#19	Anthropic	12.6%
Claude Opus 4.7	#20	Anthropic	12%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
OpenAI	7	25.8%
Moonshot AI	1	23.4%
Z.AI	1	20.9%
Anthropic	5	18.2%
Google	2	15.4%
xAI	1	15.4%
Meta	1	15.1%
Alibaba Cloud	1	13.4%
DeepSeek	1	12.9%

Benchmark score table (95 models)

One best score per model · higher is better

CritPt leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	32.3%
#2	GPT-5.5 Pro gpt.5.5.pro	OpenAI	closed	30.6%
#3	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	30%
#4	Claude Fable 5 claude-fable-5	Anthropic	closed	28.6%
#5	GPT-5.5 gpt-5.5	OpenAI	closed	27.1%
#6	GPT-5.4 gpt-5.4	OpenAI	closed	23.4%
#7	Kimi K3 kimi-k3	Moonshot AI	closed	23.4%
#8	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	20.9%
#9	GLM-5.2 glm-5-2	Z.AI	open	20.9%
#10	GPT-5.6 Luna gpt-5.6-luna	OpenAI	closed	20.6%
#11	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	17.7%
#12	Claude Sonnet 5 claude-sonnet-5	Anthropic	closed	16.9%
#13	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	16.9%
#14	Grok 4.5 grok-4.5	xAI	closed	15.4%
#15	Muse Spark 1.1 muse-spark-1-1	Meta	closed	15.1%
#16	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	13.4%
#17	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	13.1%
#18	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	12.9%
#19	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	12.6%
#20	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	12%
#21	GPT-5.2 gpt-5.2	OpenAI	closed	11.6%
#22	Muse Spark muse-spark	Meta	closed	11.3%
#23	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	10%
#24	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	10%
#25	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	9.3%
#26	Gemini 3 Pro gemini-3-pro	Google	closed	9.1%
#27	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	9.1%
#28	GPT-5.2 Codex gpt-5.2-codex	OpenAI	closed	8.7%
#29	Gemini 3 Flash gemini-3-flash	Google	closed	8.6%
#30	Grok 4.3 grok-4.3	xAI	closed	8%
#31	Kimi K2.6 kimi-k2.6	Moonshot AI	open	8%
#32	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	7.1%
#33	Grok 4.20 grok-4.20	xAI	closed	6.6%
#34	GPT-5 gpt-5	OpenAI	closed	5.7%
#35	GPT-5 Codex gpt-5-codex	OpenAI	closed	5.1%
#36	GPT-5.1 gpt.5.1	OpenAI	closed	4.9%
#37	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	4.6%
#38	GLM-5.1 glm-5-1	Z.AI	open	4.6%
#39	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	4.3%
#40	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	4%
#41	MiMo-V2.5 mimo-v2.5	Xiaomi	open	3.7%
#42	MiniMax M3 MiniMax-M3	MiniMax	—	3.7%
#43	Qwen3.6 Max qwen3.6-max	Alibaba Cloud	closed	3.7%
#44	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	3.1%
#45	Kimi K2.5 kimi-k2.5	Moonshot AI	—	3.1%
#46	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	3.1%
#47	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	3.1%
#48	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	2.9%
#49	Grok 4 Fast grok-4-fast	xAI	closed	2.9%
#50	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	2.9%
#51	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	2.6%
#52	Kimi K2 Thinking kimi-k2-thinking	Moonshot AI	open	2.6%
#53	GLM-5 glm-5	Z.AI	open	2%
#54	Grok 4 grok-4	xAI	closed	2%
#55	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	1.7%
#56	Qwen3 Max qwen3-max	Alibaba Cloud	closed	1.7%
#57	GLM-4.7 glm-4.7	Z.AI	open	1.7%
#58	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	1.7%
#59	Gemma 4 31B gemma-4-31b	Google	open	1.4%
#60	GPT-5 mini gpt-5-mini	OpenAI	closed	1.4%
#61	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	1.1%
#62	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	1.1%
#63	GLM-4.6 glm-4.6	Z.AI	open	1.1%
#64	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	1.1%
#65	MiniMax M2.5 minimax-m2.5	MiniMax	open	1.1%
#66	o3 o3	OpenAI	closed	1.1%
#67	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	1.1%
#68	Claude Sonnet 3.7 claude-3-7-sonnet	Anthropic	closed	0.9%
#69	MiniMax M2 minimax-m2	MiniMax	open	0.9%
#70	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	0.9%
#71	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	0.9%
#72	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	0.6%
#73	Grok 3 mini grok-3-mini	xAI	closed	0.6%
#74	MiniMax M2.7 minimax-m2-7	MiniMax	open	0.6%
#75	o4-mini o4-mini	OpenAI	closed	0.6%
#76	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	0.6%
#77	Qwen3.5 Omni Plus qwen3.5-omni-plus	Alibaba Cloud	closed	0.6%
#78	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	0.3%
#79	Command A+ command-a-plus	Cohere	open	0.3%
#80	Gemini 2.5 Flash gemini-2.5-flash	Google	closed	0.3%
#81	GLM-5-Turbo glm-5-turbo	Z.AI	closed	0.3%
#82	Ling 2.6 1T ling-2.6-1t	InclusionAI	open	0.3%
#83	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	0.3%
#84	MiniMax M2.1 minimax-m2.1	MiniMax	open	0.3%
#85	Mistral Small 4 mistral-small-2603	Mistral AI	open	0.3%
#86	o1 o1	OpenAI	closed	0.3%
#87	Claude Haiku 4.5 claude-haiku-4-5-20251001	Anthropic	closed	0%
#88	Claude Opus 4.1 claude-opus-4-1	Anthropic	closed	0%
#89	Gemma 4 12B gemma-4-12b	Google	open	0%
#90	Gemma 4 26B gemma-4-26b-a4b	Google	open	0%
#91	Grok Code Fast 1 grok-code-fast-1	xAI	closed	0%
#92	Kimi K2 0905 kimi-k2-0905	Moonshot AI	open	0%
#93	Mistral Large 3 mistral-large-2512	Mistral AI	open	0%
#94	Mistral Medium 3.5 mistral-medium-3-5	Mistral AI	open	0%
#95	Nova 2 Pro nova-2-pro-preview	Amazon	closed	0%

The top of this snapshot is led by GPT-5.6 Sol at 32.3%; third place is 2.3 points behind. The top-10 spread is 11.7 points.

Compare top models

GPT-5.6 Sol vs GPT-5.5 Pro GPT-5.5 Pro vs GPT-5.6 Terra GPT-5.6 Terra vs Claude Fable 5 Claude Fable 5 vs GPT-5.5

About CritPt

Definition and scoring

Organisation: CritPt authors
Category: Reasoning
Version: Not evaluated
Direction: higher is better
Ranking use: Weighted 1.5%
Contamination risk: Low

Open benchmark source ↗

FAQ

Quick answers

What does CritPt measure?

Research-level physics reasoning benchmark with composite challenges designed to stress frontier scientific reasoning.

Which model scores highest on CritPt?

GPT-5.6 Sol by OpenAI currently leads with 32.3%.

How many models are evaluated on CritPt?

95 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

Yes. This family is ranking-weighted at 1.5% of its capability category. GPT-5.6 Sol is overall #4.

GPQA Diamond

Rolling · 96 results · Weighted 2.5%

Humanity's Last Exam

Rolling · 105 results · Weighted 2.5%

ARC-AGI-2

2 · 10 results · Weighted 2.5%

LiveBench

Rolling · 17 results · Weighted 2.5%

ARC-AGI-3

3 · 3 results · Not Weighted

BenchLM Reasoning prior

bench-align-v5.1 · 18 results · Weighted 1.1%

Leaderboard · 95 models

GPT-5.6 Sol

GPT-5.5 Pro

GPT-5.6 Terra

See the field, not just the table.

Benchmark score table (95 models)

Compare top models

Definition and scoring

Quick answers

What does CritPt measure?

Which model scores highest on CritPt?

How many models are evaluated on CritPt?

Does this affect overall Lumina rank?

More Reasoning benchmarks

Leaderboard · 95 models

GPT-5.6 Sol

GPT-5.5 Pro

GPT-5.6 Terra

See the field, not just the table.

Benchmark score table (95 models)

Compare top models

Definition and scoring

Quick answers

What does CritPt measure?

Which model scores highest on CritPt?

How many models are evaluated on CritPt?

Does this affect overall Lumina rank?

More Reasoning benchmarks