Reasoning · Benchmark profile

Humanity's Last Exam

A broad expert-level academic benchmark covering difficult questions across many domains.

Ranking weightedCurrent / rollingRolling88 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on Humanity's Last Exam

Leaderboard · 88 models

Claude Mythos 5 leads at 64.5%, followed by Muse Spark 1.1 (62.1%) and Claude Opus 4.8 (57.9%).

Ranking · 2.5%ReasoningCurrent

#1Closed

Claude Mythos 5

Anthropic

claude-mythos-5

64.5%

Overall #1 · Context 1M tokens

#2Closed

Claude Opus 4.8

Anthropic

claude-opus-4-8

57.9%

Overall #5 · Context 1M 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

AnthropicMetaOpenAIMoonshot AIZ.AITencentXiaomiGoogle

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Claude Mythos 5	#1	Anthropic	64.5%
Muse Spark 1.1	#2	Meta	62.1%
Claude Opus 4.8	#3	Anthropic	57.9%
Claude Sonnet 5	#4	Anthropic	57.4%
GPT-5.5 Pro	#5	OpenAI	57.2%
Kimi K3	#6	Moonshot AI	56%
GLM-5.2	#7	Z.AI	54.7%
Claude Fable 5	#8	Anthropic	53.3%
Hy3	#9	Tencent	53.2%
Claude Opus 4.6	#10	Anthropic	53%
GLM-5.1	#11	Z.AI	52.3%
GPT-5.5	#12	OpenAI	52.2%
GPT-5.4	#13	OpenAI	52.1%
GLM-5	#14	Z.AI	50.4%
Muse Spark	#15	Meta	50.4%
Claude Sonnet 4.6	#16	Anthropic	49%
MiMo-V2.5-Pro	#17	Xiaomi	48%
GPT-5.6 Sol	#18	OpenAI	47.2%
Gemini 3.1 Pro Preview	#19	Google	44.7%
GPT-5.6 Terra	#20	OpenAI	41.8%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Claude Mythos 5	#1	Anthropic	64.5%
Muse Spark 1.1	#2	Meta	62.1%
Claude Opus 4.8	#3	Anthropic	57.9%
Claude Sonnet 5	#4	Anthropic	57.4%
GPT-5.5 Pro	#5	OpenAI	57.2%
Kimi K3	#6	Moonshot AI	56%
GLM-5.2	#7	Z.AI	54.7%
Claude Fable 5	#8	Anthropic	53.3%
Hy3	#9	Tencent	53.2%
Claude Opus 4.6	#10	Anthropic	53%
GLM-5.1	#11	Z.AI	52.3%
GPT-5.5	#12	OpenAI	52.2%
GPT-5.4	#13	OpenAI	52.1%
GLM-5	#14	Z.AI	50.4%
Muse Spark	#15	Meta	50.4%
Claude Sonnet 4.6	#16	Anthropic	49%
MiMo-V2.5-Pro	#17	Xiaomi	48%
GPT-5.6 Sol	#18	OpenAI	47.2%
Gemini 3.1 Pro Preview	#19	Google	44.7%
GPT-5.6 Terra	#20	OpenAI	41.8%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
Meta	2	56.3%
Moonshot AI	1	56%
Anthropic	6	55.9%
Tencent	1	53.2%
Z.AI	3	52.5%
OpenAI	5	50.1%
Xiaomi	1	48%
Google	1	44.7%

Benchmark score table (88 models)

One best score per model · higher is better

Humanity's Last Exam leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Claude Mythos 5 claude-mythos-5	Anthropic	closed	64.5%
#2	Muse Spark 1.1 muse-spark-1-1	Meta	closed	62.1%
#3	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	57.9%
#4	Claude Sonnet 5 claude-sonnet-5	Anthropic	closed	57.4%
#5	GPT-5.5 Pro gpt.5.5.pro	OpenAI	closed	57.2%
#6	Kimi K3 kimi-k3	Moonshot AI	closed	56%
#7	GLM-5.2 glm-5-2	Z.AI	open	54.7%
#8	Claude Fable 5 claude-fable-5	Anthropic	closed	53.3%
#9	Hy3 hy3	Tencent	open	53.2%
#10	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	53%
#11	GLM-5.1 glm-5-1	Z.AI	open	52.3%
#12	GPT-5.5 gpt-5.5	OpenAI	closed	52.2%
#13	GPT-5.4 gpt-5.4	OpenAI	closed	52.1%
#14	GLM-5 glm-5	Z.AI	open	50.4%
#15	Muse Spark muse-spark	Meta	closed	50.4%
#16	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	49%
#17	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	48%
#18	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	47.2%
#19	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	44.7%
#20	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	41.8%
#21	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	41.4%
#22	Grok 4.5 grok-4.5	xAI	closed	40.3%
#23	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	40.2%
#24	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	39.9%
#25	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	37.7%
#26	GPT-5.6 Luna gpt-5.6-luna	OpenAI	closed	37.2%
#27	MiniMax M3 MiniMax-M3	MiniMax	—	37.1%
#28	Kimi K2.6 kimi-k2.6	Moonshot AI	open	35.9%
#29	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	35.9%
#30	GPT-5.2 gpt-5.2	OpenAI	closed	35.4%
#31	Grok 4.3 grok-4.3	xAI	closed	35.0%
#32	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	34.7%
#33	GPT-5.2 Codex gpt-5.2-codex	OpenAI	closed	33.5%
#34	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	32.8%
#35	Grok 4.20 grok-4.20	xAI	closed	32.2%
#36	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	32.1%
#37	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	30.8%
#38	Kimi K2.5 kimi-k2.5	Moonshot AI	—	30.1%
#39	Qwen3.6 Max qwen3.6-max	Alibaba Cloud	closed	28.9%
#40	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	28.8%
#41	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	27.3%
#42	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	26.6%
#43	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	26.6%
#44	GPT-5 gpt-5	OpenAI	closed	26.5%
#45	Gemma 4 31B gemma-4-31b	Google	open	26.5%
#46	Qwen3 Max qwen3-max	Alibaba Cloud	closed	26.2%
#47	GPT-5 Codex gpt-5-codex	OpenAI	closed	25.6%
#48	MiMo-V2.5 mimo-v2.5	Xiaomi	open	25.2%
#49	GLM-4.7 glm-4.7	Z.AI	open	25.1%
#50	Grok 4 grok-4	xAI	closed	23.9%
#51	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	23.4%
#52	Kimi K2 Thinking kimi-k2-thinking	Moonshot AI	open	22.3%
#53	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	22.2%
#54	MiniMax M2.1 minimax-m2.1	MiniMax	open	22.2%
#55	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	22.2%
#56	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	21.6%
#57	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	21.1%
#58	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	21.1%
#59	o3 o3	OpenAI	closed	20.0%
#60	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	19.7%
#61	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	19.2%
#62	MiniMax M2.5 minimax-m2.5	MiniMax	open	19.1%
#63	Gemma 4 26B gemma-4-26b-a4b	Google	open	18.3%
#64	o4-mini o4-mini	OpenAI	closed	17.5%
#65	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	17.3%
#66	Grok 4 Fast grok-4-fast	xAI	closed	17.0%
#67	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	15.2%
#68	Gemma 4 12B gemma-4-12b	Google	open	14.8%
#69	GPT-5 mini gpt-5-mini	OpenAI	closed	14.6%
#70	Qwen3.5 Omni Plus qwen3.5-omni-plus	Alibaba Cloud	closed	13.9%
#71	GLM-4.6 glm-4.6	Z.AI	open	13.3%
#72	Mistral Medium 3.5 mistral-medium-3-5	Mistral AI	open	12.8%
#73	Gemini 2.5 Flash gemini-2.5-flash	Google	closed	12.7%
#74	MiniMax M2 minimax-m2	MiniMax	open	12.5%
#75	Claude Opus 4.1 claude-opus-4-1	Anthropic	closed	11.9%
#76	Claude Opus 4 claude-opus-4	Anthropic	closed	11.7%
#77	Command A+ command-a-plus	Cohere	open	11.4%
#78	Grok 3 mini grok-3-mini	xAI	closed	11.1%
#79	Claude Sonnet 3.7 claude-3-7-sonnet	Anthropic	closed	10.3%
#80	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	9.6%
#81	Mistral Small 4 mistral-small-2603	Mistral AI	open	9.5%
#82	Nova 2 Pro nova-2-pro-preview	Amazon	closed	8.9%
#83	Ling 2.6 1T ling-2.6-1t	InclusionAI	open	8.2%
#84	o1 o1	OpenAI	closed	7.7%
#85	Grok Code Fast 1 grok-code-fast-1	xAI	closed	7.5%
#86	Kimi K2 0905 kimi-k2-0905	Moonshot AI	open	6.3%
#87	Claude Haiku 4.5 claude-haiku-4-5-20251001	Anthropic	closed	4.3%
#88	Mistral Large 3 mistral-large-2512	Mistral AI	open	4.1%

The top of this snapshot is led by Claude Mythos 5 at 64.5%; third place is 6.6 points behind. The top-10 spread is 11.5 points.

Compare top models

Claude Mythos 5 vs Muse Spark 1.1 Muse Spark 1.1 vs Claude Opus 4.8 Claude Opus 4.8 vs Claude Sonnet 5 Claude Sonnet 5 vs GPT-5.5 Pro

About Humanity's Last Exam

Definition and scoring

Organisation: Center for AI Safety and Scale AI
Category: Reasoning
Version: Not evaluated
Direction: higher is better
Ranking use: Weighted 2.5%
Contamination risk: Unknown

A broad expert-level academic benchmark covering difficult questions across many domains. 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 Humanity's Last Exam measure?

A broad expert-level academic benchmark covering difficult questions across many domains.

Which model scores highest on Humanity's Last Exam?

Claude Mythos 5 by Anthropic currently leads with 64.5%.

How many models are evaluated on Humanity's Last Exam?

88 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 2.5% of its capability category. Claude Mythos 5 is overall #1.

GPQA Diamond

Rolling · 96 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%

SuperGPQA

Rolling · 8 results · Weighted 2.5%

Reasoning · Benchmark profile

Humanity's Last Exam

A broad expert-level academic benchmark covering difficult questions across many domains.

Ranking weightedCurrent / rollingRolling88 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on Humanity's Last Exam

Leaderboard · 88 models

Claude Mythos 5 leads at 64.5%, followed by Muse Spark 1.1 (62.1%) and Claude Opus 4.8 (57.9%).

Ranking · 2.5%ReasoningCurrent

#1Closed

Claude Mythos 5

Anthropic

claude-mythos-5

64.5%

Overall #1 · Context 1M tokens

#2Closed

Muse Spark 1.1

Claude Opus 4.8

Anthropic

claude-opus-4-8

57.9%

Overall #5 · Context 1M 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

AnthropicMetaOpenAIMoonshot AIZ.AITencentXiaomiGoogle

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Claude Mythos 5	#1	Anthropic	64.5%
Muse Spark 1.1	#2	Meta	62.1%
Claude Opus 4.8	#3	Anthropic	57.9%
Claude Sonnet 5	#4	Anthropic	57.4%
GPT-5.5 Pro	#5	OpenAI	57.2%
Kimi K3	#6	Moonshot AI	56%
GLM-5.2	#7	Z.AI	54.7%
Claude Fable 5	#8	Anthropic	53.3%
Hy3	#9	Tencent	53.2%
Claude Opus 4.6	#10	Anthropic	53%
GLM-5.1	#11	Z.AI	52.3%
GPT-5.5	#12	OpenAI	52.2%
GPT-5.4	#13	OpenAI	52.1%
GLM-5	#14	Z.AI	50.4%
Muse Spark	#15	Meta	50.4%
Claude Sonnet 4.6	#16	Anthropic	49%
MiMo-V2.5-Pro	#17	Xiaomi	48%
GPT-5.6 Sol	#18	OpenAI	47.2%
Gemini 3.1 Pro Preview	#19	Google	44.7%
GPT-5.6 Terra	#20	OpenAI	41.8%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Claude Mythos 5	#1	Anthropic	64.5%
Muse Spark 1.1	#2	Meta	62.1%
Claude Opus 4.8	#3	Anthropic	57.9%
Claude Sonnet 5	#4	Anthropic	57.4%
GPT-5.5 Pro	#5	OpenAI	57.2%
Kimi K3	#6	Moonshot AI	56%
GLM-5.2	#7	Z.AI	54.7%
Claude Fable 5	#8	Anthropic	53.3%
Hy3	#9	Tencent	53.2%
Claude Opus 4.6	#10	Anthropic	53%
GLM-5.1	#11	Z.AI	52.3%
GPT-5.5	#12	OpenAI	52.2%
GPT-5.4	#13	OpenAI	52.1%
GLM-5	#14	Z.AI	50.4%
Muse Spark	#15	Meta	50.4%
Claude Sonnet 4.6	#16	Anthropic	49%
MiMo-V2.5-Pro	#17	Xiaomi	48%
GPT-5.6 Sol	#18	OpenAI	47.2%
Gemini 3.1 Pro Preview	#19	Google	44.7%
GPT-5.6 Terra	#20	OpenAI	41.8%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
Meta	2	56.3%
Moonshot AI	1	56%
Anthropic	6	55.9%
Tencent	1	53.2%
Z.AI	3	52.5%
OpenAI	5	50.1%
Xiaomi	1	48%
Google	1	44.7%

Benchmark score table (88 models)

One best score per model · higher is better

Humanity's Last Exam leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Claude Mythos 5 claude-mythos-5	Anthropic	closed	64.5%
#2	Muse Spark 1.1 muse-spark-1-1	Meta	closed	62.1%
#3	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	57.9%
#4	Claude Sonnet 5 claude-sonnet-5	Anthropic	closed	57.4%
#5	GPT-5.5 Pro gpt.5.5.pro	OpenAI	closed	57.2%
#6	Kimi K3 kimi-k3	Moonshot AI	closed	56%
#7	GLM-5.2 glm-5-2	Z.AI	open	54.7%
#8	Claude Fable 5 claude-fable-5	Anthropic	closed	53.3%
#9	Hy3 hy3	Tencent	open	53.2%
#10	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	53%
#11	GLM-5.1 glm-5-1	Z.AI	open	52.3%
#12	GPT-5.5 gpt-5.5	OpenAI	closed	52.2%
#13	GPT-5.4 gpt-5.4	OpenAI	closed	52.1%
#14	GLM-5 glm-5	Z.AI	open	50.4%
#15	Muse Spark muse-spark	Meta	closed	50.4%
#16	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	49%
#17	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	48%
#18	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	47.2%
#19	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	44.7%
#20	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	41.8%
#21	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	41.4%
#22	Grok 4.5 grok-4.5	xAI	closed	40.3%
#23	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	40.2%
#24	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	39.9%
#25	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	37.7%
#26	GPT-5.6 Luna gpt-5.6-luna	OpenAI	closed	37.2%
#27	MiniMax M3 MiniMax-M3	MiniMax	—	37.1%
#28	Kimi K2.6 kimi-k2.6	Moonshot AI	open	35.9%
#29	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	35.9%
#30	GPT-5.2 gpt-5.2	OpenAI	closed	35.4%
#31	Grok 4.3 grok-4.3	xAI	closed	35.0%
#32	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	34.7%
#33	GPT-5.2 Codex gpt-5.2-codex	OpenAI	closed	33.5%
#34	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	32.8%
#35	Grok 4.20 grok-4.20	xAI	closed	32.2%
#36	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	32.1%
#37	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	30.8%
#38	Kimi K2.5 kimi-k2.5	Moonshot AI	—	30.1%
#39	Qwen3.6 Max qwen3.6-max	Alibaba Cloud	closed	28.9%
#40	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	28.8%
#41	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	27.3%
#42	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	26.6%
#43	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	26.6%
#44	GPT-5 gpt-5	OpenAI	closed	26.5%
#45	Gemma 4 31B gemma-4-31b	Google	open	26.5%
#46	Qwen3 Max qwen3-max	Alibaba Cloud	closed	26.2%
#47	GPT-5 Codex gpt-5-codex	OpenAI	closed	25.6%
#48	MiMo-V2.5 mimo-v2.5	Xiaomi	open	25.2%
#49	GLM-4.7 glm-4.7	Z.AI	open	25.1%
#50	Grok 4 grok-4	xAI	closed	23.9%
#51	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	23.4%
#52	Kimi K2 Thinking kimi-k2-thinking	Moonshot AI	open	22.3%
#53	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	22.2%
#54	MiniMax M2.1 minimax-m2.1	MiniMax	open	22.2%
#55	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	22.2%
#56	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	21.6%
#57	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	21.1%
#58	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	21.1%
#59	o3 o3	OpenAI	closed	20.0%
#60	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	19.7%
#61	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	19.2%
#62	MiniMax M2.5 minimax-m2.5	MiniMax	open	19.1%
#63	Gemma 4 26B gemma-4-26b-a4b	Google	open	18.3%
#64	o4-mini o4-mini	OpenAI	closed	17.5%
#65	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	17.3%
#66	Grok 4 Fast grok-4-fast	xAI	closed	17.0%
#67	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	15.2%
#68	Gemma 4 12B gemma-4-12b	Google	open	14.8%
#69	GPT-5 mini gpt-5-mini	OpenAI	closed	14.6%
#70	Qwen3.5 Omni Plus qwen3.5-omni-plus	Alibaba Cloud	closed	13.9%
#71	GLM-4.6 glm-4.6	Z.AI	open	13.3%
#72	Mistral Medium 3.5 mistral-medium-3-5	Mistral AI	open	12.8%
#73	Gemini 2.5 Flash gemini-2.5-flash	Google	closed	12.7%
#74	MiniMax M2 minimax-m2	MiniMax	open	12.5%
#75	Claude Opus 4.1 claude-opus-4-1	Anthropic	closed	11.9%
#76	Claude Opus 4 claude-opus-4	Anthropic	closed	11.7%
#77	Command A+ command-a-plus	Cohere	open	11.4%
#78	Grok 3 mini grok-3-mini	xAI	closed	11.1%
#79	Claude Sonnet 3.7 claude-3-7-sonnet	Anthropic	closed	10.3%
#80	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	9.6%
#81	Mistral Small 4 mistral-small-2603	Mistral AI	open	9.5%
#82	Nova 2 Pro nova-2-pro-preview	Amazon	closed	8.9%
#83	Ling 2.6 1T ling-2.6-1t	InclusionAI	open	8.2%
#84	o1 o1	OpenAI	closed	7.7%
#85	Grok Code Fast 1 grok-code-fast-1	xAI	closed	7.5%
#86	Kimi K2 0905 kimi-k2-0905	Moonshot AI	open	6.3%
#87	Claude Haiku 4.5 claude-haiku-4-5-20251001	Anthropic	closed	4.3%
#88	Mistral Large 3 mistral-large-2512	Mistral AI	open	4.1%

The top of this snapshot is led by Claude Mythos 5 at 64.5%; third place is 6.6 points behind. The top-10 spread is 11.5 points.

Compare top models

Claude Mythos 5 vs Muse Spark 1.1 Muse Spark 1.1 vs Claude Opus 4.8 Claude Opus 4.8 vs Claude Sonnet 5 Claude Sonnet 5 vs GPT-5.5 Pro

About Humanity's Last Exam

Definition and scoring

Organisation: Center for AI Safety and Scale AI
Category: Reasoning
Version: Not evaluated
Direction: higher is better
Ranking use: Weighted 2.5%
Contamination risk: Unknown

Open benchmark source ↗

FAQ

Quick answers

What does Humanity's Last Exam measure?

A broad expert-level academic benchmark covering difficult questions across many domains.

Which model scores highest on Humanity's Last Exam?

Claude Mythos 5 by Anthropic currently leads with 64.5%.

How many models are evaluated on Humanity's Last Exam?

88 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 2.5% of its capability category. Claude Mythos 5 is overall #1.

GPQA Diamond

Rolling · 96 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%

SuperGPQA

Rolling · 8 results · Weighted 2.5%

Leaderboard · 88 models

Claude Mythos 5

Muse Spark 1.1

Claude Opus 4.8

See the field, not just the table.

Benchmark score table (88 models)

Compare top models

Definition and scoring

Quick answers

What does Humanity's Last Exam measure?

Which model scores highest on Humanity's Last Exam?

How many models are evaluated on Humanity's Last Exam?

Does this affect overall Lumina rank?

More Reasoning benchmarks

Leaderboard · 88 models

Claude Mythos 5

Muse Spark 1.1

Claude Opus 4.8

See the field, not just the table.

Benchmark score table (88 models)

Compare top models

Definition and scoring

Quick answers

What does Humanity's Last Exam measure?

Which model scores highest on Humanity's Last Exam?

How many models are evaluated on Humanity's Last Exam?

Does this affect overall Lumina rank?

More Reasoning benchmarks