BBH Leaderboard

Big-Bench Hard — challenging subset of BIG-Bench focusing on tasks where language models previously underperformed.

As of April 4, 2026, the top-scoring model on BBH is Claude Sonnet 4.5 at 94.4%, followed by Claude Sonnet 4.5 at 94.3% and GLM 5 at 94.3%. 109 models have been evaluated on this benchmark.

Last updated: April 4, 2026

Models

109

Best Score

94.4

Average

80.7

Std Dev

13.3

Categories

Reasoning and Logic

SourceLayerLens

Provider	Model	Input $/M	Output $/M	BBH	Actions
A Anthropic	Claude Sonnet 4.5 Thinking	$3.000	$15.000	94.4	Try
A Anthropic	Claude Sonnet 4.5	$3.000	$15.000	94.3	Try
Z Z AI	GLM 5 Thinking	$0.720	$2.300	94.3	Try
Z Z AI	GLM 5	$0.720	$2.300	94.3	Try
G Google	Gemini 3 Pro Preview	$2.000	$12.000	93.8	Try
AL Alibaba	Qwen3.5-122B-A10B	$0.260	$2.080	92.7	Try
AL Alibaba	Qwen3.5-122B-A10B	$0.260	$2.080	92.7	Try
AL Alibaba	Qwen3 Max	$0.780	$3.900	92.6	Try
AL Alibaba	Qwen3 Max	$0.780	$3.900	92.6	Try
AL Alibaba	Qwen3.5-27B	$0.195	$0.900	92.4	Try
AL Alibaba	Qwen3.5-27B	$0.195	$0.900	92.4	Try
K Kimi	Kimi K2.5 Thinking	$0.383	$1.720	91.0	Try
K Kimi	Kimi K2.5	$0.383	$1.720	91.0	Try
A Anthropic	Claude 3.7 Sonnet Thinking	$3.000	$15.000	91.0	Try
A Anthropic	Claude 3.7 Sonnet	$3.000	$15.000	91.0	Try
O OpenAI	GPT-5	$0.625	$5.000	90.9	Try
O OpenAI	GPT-5	$0.625	$5.000	90.9	Try
O OpenAI	GPT-5	$0.625	$5.000	90.9	Try
O OpenAI	GPT-5	$0.625	$5.000	90.9	Try
AL Alibaba	Qwen3.5-35B-A3B	$0.163	$0.900	90.8	Try
AL Alibaba	Qwen3.5-35B-A3B	$0.163	$0.900	90.8	Try
QW Qwen	Qwen3.5-Flash	$0.065	$0.260	90.6	Try
QW Qwen	Qwen3.5 Plus	$0.260	$1.560	90.6	Try
A Anthropic	Claude Sonnet 4.6	$3.000	$15.000	90.3	Try
A Anthropic	Claude Sonnet 4.6	$3.000	$15.000	90.3	Try
A Anthropic	Claude Sonnet 4.6	$3.000	$15.000	90.3	Try
X xAI	Grok 4 Fast	$0.200	$0.500	89.6	Try
O OpenAI	o3	$2.000	$8.000	89.6	Try
AL Alibaba	Qwen3 30B A3B Thinking	$0.080	$0.280	89.5	Try
AL Alibaba	Qwen3 30B A3B	$0.080	$0.280	89.5	Try
AL Alibaba	Qwen3.5 397B A17B	$0.390	$0.900	89.4	Try
AL Alibaba	Qwen3.5 397B A17B	$0.390	$0.900	89.4	Try
X xAI	Grok 4 Fast Thinking	$0.200	$0.500	89.3	Try
M Meta	Llama 3.1 405B Instruct	$0.900	$0.900	89.0	Try
G Google	Gemini 2.5 Pro	$1.000	$10.000	88.7	Try
DS DeepSeek	DeepSeek V3.2 Thinking	$0.260	$0.380	88.5	Try
Z Z AI	GLM 4.7 Thinking	$0.390	$1.750	88.4	Try
Z Z AI	GLM 4.7	$0.390	$1.750	88.4	Try
G Google	Gemini 2.0 Flash	$0.100	$0.400	88.3	Try
DS DeepSeek	DeepSeek V3.2 Exp Thinking	$0.270	$0.410	88.0	Try
DS DeepSeek	R1 0528	$0.450	$2.150	87.9	Try
O OpenAI	GPT-5 Mini	$0.125	$1.000	87.7	Try
AL Alibaba	Qwen3 235B A22B Instruct 2507	$0.071	$0.100	87.4	Try
K Kimi	Kimi K2 0711	$0.550	$2.200	87.1	Try
K Kimi	Kimi K2 0711	$0.550	$2.200	87.1	Try
O OpenAI	GPT-5 Nano	$0.050	$0.400	86.9	Try
O OpenAI	GPT-5 Nano	$0.050	$0.400	86.9	Try
O OpenAI	GPT-5 Nano	$0.050	$0.400	86.9	Try
K Kimi	Kimi K2 0905 (exacto)	$0.400	$2.000	86.7	Try
AL Alibaba	Qwen3 32B Thinking	$0.080	$0.240	85.8	Try
AL Alibaba	Qwen3 32B	$0.080	$0.240	85.8	Try
BD Baidu	ERNIE 4.5 300B A47B	$0.280	$0.900	85.7	Try
Z Z AI	GLM 4.6 Thinking	$0.390	$1.700	85.5	Try
Z Z AI	GLM 4.6	$0.390	$1.700	85.5	Try
MM MiniMax	MiniMax M2.1	$0.270	$0.950	85.0	Try
AL Alibaba	Qwen3 14B Thinking	$0.060	$0.200	84.6	Try
AL Alibaba	Qwen3 14B	$0.060	$0.200	84.6	Try
MM MiniMax	MiniMax M2.5	$0.118	$0.950	84.5	Try
O OpenAI	o3 Mini	$1.100	$4.400	84.4	Try
MM MiniMax	MiniMax M2	$0.255	$1.000	84.4	Try
MS Moonshotai	Kimi K2 Thinking	$0.470	$2.000	84.4	Try
AL Alibaba	Qwen3 235B A22B Thinking	$0.400	$0.800	84.2	Try
AL Alibaba	Qwen3 235B A22B	$0.400	$0.800	84.2	Try
O OpenAI	GPT-OSS-20b	$0.030	$0.100	84.1	Try
AM Amazon	Nova Pro 1.0	$0.800	$3.200	83.9	Try
Z Z AI	GLM 4.5 Thinking	$0.600	$2.200	83.8	Try
DS DeepSeek	R1	$0.550	$2.000	83.5	Try
O OpenAI	GPT-OSS-120b	$0.039	$0.100	82.1	Try
O OpenAI	GPT-OSS-120b	$0.039	$0.100	82.1	Try
O OpenAI	o4 Mini High	$1.100	$4.400	82.0	Try
A Anthropic	Claude Haiku 4.5 Thinking	$1.000	$5.000	82.0	Try
A Anthropic	Claude Haiku 4.5	$1.000	$5.000	82.0	Try
M Meta	Llama 4 Maverick	$0.150	$0.600	81.8	Try
MI Mistral	Mistral Medium 3	$0.400	$2.000	81.5	Try
AL Alibaba	QwQ 32B	$0.150	$0.580	81.2	Try
AL Alibaba	QwQ 32B	$0.150	$0.580	81.2	Try
MI Mistral	Mistral Medium 3.1	$0.400	$2.000	80.5	Try
CO Cohere	Command A	$2.500	$10.000	79.6	Try
M Meta	Llama 4 Scout	$0.080	$0.300	79.3	Try
Z Z AI	GLM 4.5 Air	$0.130	$0.850	77.9	Try
A Anthropic	Claude 3.5 Haiku	$0.800	$4.000	77.0	Try
G Google	Gemma 3 27B	$0.080	$0.160	76.6	Try
MI Mistral	Mistral Small 3.2 24B	$0.075	$0.200	76.5	Try
MM MiniMax	MiniMax M1	$0.400	$1.760	74.3	Try
G Google	Gemma 3n 4B	$0.020	$0.040	72.7	Try
AL Alibaba	Qwen2.5 72B Instruct	$0.120	$0.390	72.4	Try
X xAI	Grok 4.1 Fast Thinking	$0.200	$0.500	72.1	Try
X xAI	Grok 4.1 Fast	$0.200	$0.500	71.6	Try
MI Mistral	Mistral Large 2411	$2.000	$6.000	70.3	Try
DS Deepseek	DeepSeek V3	$0.014	$0.028	69.8	Try
O OpenAI	GPT-5.1	$0.625	$5.000	69.8	Try
O OpenAI	GPT-5.1	$0.625	$5.000	69.8	Try
O OpenAI	GPT-5.2	$1.750	$14.000	67.6	Try
MI Mistral	Devstral Small 1.1	$0.070	$0.280	64.8	Try
NO Nousresearch	Hermes 3 405B Instruct	$1.000	$1.000	64.2	Try
G Google	Gemini 2.5 Flash Thinking	$0.300	$2.500	61.6	Try
G Google	Gemini 2.5 Flash	$0.300	$2.500	61.6	Try
AM Amazon	Nova Micro 1.0	$0.035	$0.140	61.5	Try
DS DeepSeek	DeepSeek V3.1 Terminus Thinking	$0.210	$0.790	60.5	Try
DS DeepSeek	DeepSeek V3.1 Terminus	$0.210	$0.790	60.5	Try
AM Amazon	Nova Lite 1.0	$0.060	$0.240	59.0	Try
X Xai	Grok 3 Beta	$3.000	$15.000	53.8	Try
X xAI	Grok 3	$3.000	$15.000	53.4	Try
X xAI	Grok 3	$3.000	$15.000	53.4	Try
DS DeepSeek	DeepSeek V3.2 Exp	$0.270	$0.410	52.8	Try
MI Mistral	Devstral Medium	$0.400	$2.000	48.1	Try
MS Microsoft Azure	Phi 4	$0.065	$0.140	39.4	Try
M Meta	Llama 3.2 3B Instruct	$0.030	$0.050	36.0	Try
IF Inflection	Inflection 3 Pi	$2.500	$10.000	28.2	Try

Pricing from OpenRouter. Benchmarks from Artificial Analysis.

Get our weekly newsletter on pricing changes, new releases, and tools.

Deploy OpenClaw in Under 1 Minute— We handle hosting, scaling, and maintenance

8 Ways to Use Fewer Tokens

About BBH

Big-Bench Hard — challenging subset of BIG-Bench focusing on tasks where language models previously underperformed.

This leaderboard shows all models with BBH benchmark scores, ranked from highest to lowest. Pricing data is included to help you compare performance against cost.