math benchmark
BBH
Big-Bench Hard (BBH) is a suite of 23 challenging tasks selected from BIG-Bench for which prior language model evaluations did not outperform the average human-rater. These tasks require multi-step reasoning across diverse domains including arithmetic, logical reasoning, reading comprehension, and commonsense reasoning. The benchmark was designed to test capabilities believed to be beyond current language models and focuses on evaluating complex reasoning skills including temporal understanding, spatial reasoning, causal understanding, and deductive logical reasoning.
Updated Aug 7, 2026
Published models12
Registry coverage12
MetricScore
EvidenceB
BBH leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 88.9% | 100.0% | 12 | C | |
| 2 | XI | 88.4% | 90.9% | 12 | C | |
| 3 | AM Nova ProAmazon | 86.9% | 81.8% | 12 | C | |
| 4 | AC | 84.5% | 72.7% | 12 | C | |
| 5 | DE | 84.3% | 63.6% | 12 | C | |
| 6 | AMNova LiteAmazon | 82.4% | 54.5% | 12 | C | |
| 7 | AC | 82.4% | 45.5% | 12 | C | |
| 8 | OP | 81.5% | 36.4% | 12 | C | |
| 9 | AMNova MicroAmazon | 79.5% | 27.3% | 12 | C | |
| 10 | AC | 78.2% | 18.2% | 12 | C | |
| 11 | NR | 67.8% | 9.1% | 12 | C | |
| 12 | BA | 30.4% | 0.0% | 12 | C |
Score distribution
Top published rows on the benchmark's original scale.
BBH highlights
The top published results on this benchmark's own scale.
What is BBH?
Definition and scoring fields from the benchmark registry.
Big-Bench Hard (BBH) is a suite of 23 challenging tasks selected from BIG-Bench for which prior language model evaluations did not outperform the average human-rater. These tasks require multi-step reasoning across diverse domains including arithmetic, logical reasoning, reading comprehension, and commonsense reasoning. The benchmark was designed to test capabilities believed to be beyond current language models and focuses on evaluating complex reasoning skills including temporal understanding, spatial reasoning, causal understanding, and deductive logical reasoning.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- BBH
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- bbh|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about BBH.
Which model scores highest on BBH?
Qwen3 235B A22B is currently ranked first with 88.9%.
What does BBH measure?
Big-Bench Hard (BBH) is a suite of 23 challenging tasks selected from BIG-Bench for which prior language model evaluations did not outperform the average human-rater. These tasks require multi-step reasoning across diverse domains including arithmetic, logical reasoning, reading comprehension, and commonsense reasoning. The benchmark was designed to test capabilities believed to be beyond current language models and focuses on evaluating complex reasoning skills including temporal understanding, spatial reasoning, causal understanding, and deductive logical reasoning.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
12 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is marked as eligible for the current LLMBoard capability methodology.