math benchmark
HiddenMath Leaderboard
Google DeepMind's internal mathematical reasoning benchmark that introduces novel problems not encountered during model training to evaluate true mathematical reasoning capabilities rather than memorization
Updated Aug 17, 2026
Models13
Model coverage13
MetricScore
EvidenceB
HiddenMath Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelGO | Score63.0% | Percentile100.0% | Participants13 | EvidenceC | Evaluated |
| Rank02 | ModelGO | Score60.3% | Percentile91.7% | Participants13 | EvidenceC | Evaluated |
| Rank03 | ModelGO | Score55.3% | Percentile83.3% | Participants13 | EvidenceC | Evaluated |
| Rank04 | ModelGO | Score54.5% | Percentile75.0% | Participants13 | EvidenceC | Evaluated |
| Rank05 | ModelGO | Score52.0% | Percentile66.7% | Participants13 | EvidenceC | Evaluated |
| Rank06 | ModelGO | Score47.2% | Percentile58.3% | Participants13 | EvidenceC | Evaluated |
| Rank07 | ModelGO | Score43.0% | Percentile50.0% | Participants13 | EvidenceC | Evaluated |
| Rank08 | ModelGO | Score37.7% | Percentile41.7% | Participants13 | EvidenceC | Evaluated |
| Rank09 | ModelGO | Score37.7% | Percentile33.3% | Participants13 | EvidenceC | Evaluated |
| Rank10 | ModelGO | Score32.8% | Percentile25.0% | Participants13 | EvidenceC | Evaluated |
| Rank11 | ModelGO | Score27.7% | Percentile16.7% | Participants13 | EvidenceC | Evaluated |
| Rank12 | ModelGO | Score27.7% | Percentile8.3% | Participants13 | EvidenceC | Evaluated |
| Rank13 | ModelGO | Score15.8% | Percentile0.0% | Participants13 | EvidenceC | Evaluated |
HiddenMath Highlights
The leading models and scores on this benchmark.
HiddenMath Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for HiddenMath
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis hiddenmath AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01GOGemini 2.0 FlashGoogle
- Score
- 63.0%
- Speed
- Up to 183 tok/s via Google
Strengths
- Ranks #1 of 13 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures HiddenMath, not total model capability
- 02GOGemma 3 27BGoogle
- Score
- 60.3%
- Speed
- Up to 33 tok/s via DeepInfra
Strengths
- Ranks #2 of 13 compared models
- 92th percentile on this benchmark
- C evidence result
Considerations
- This result measures HiddenMath, not total model capability
- 03GOGemini 2.0 Flash-LiteGoogle
- Score
- 55.3%
- Speed
- Up to 85 tok/s via Google
Strengths
- Ranks #3 of 13 compared models
- 83th percentile on this benchmark
- C evidence result
Considerations
- This result measures HiddenMath, not total model capability
- 04GOGemma 3 12BGoogle
- Score
- 54.5%
- Speed
- Up to 33 tok/s via DeepInfra
Strengths
- Ranks #4 of 13 compared models
- 75th percentile on this benchmark
- C evidence result
Considerations
- This result measures HiddenMath, not total model capability
- 05GOGemini 1.5 ProGoogle
- Score
- 52.0%
- Speed
- Up to 85 tok/s via Google
Strengths
- Ranks #5 of 13 compared models
- 67th percentile on this benchmark
- C evidence result
Considerations
- This result measures HiddenMath, not total model capability
Selection summary
Best AI Models for HiddenMath
Gemini 2.0 Flash currently leads HiddenMath with 63.0%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is HiddenMath?
What HiddenMath measures and how its scores work.
Google DeepMind's internal mathematical reasoning benchmark that introduces novel problems not encountered during model training to evaluate true mathematical reasoning capabilities rather than memorization
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- HiddenMath
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- hiddenmath|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about HiddenMath.
Which model scores highest on HiddenMath?
Gemini 2.0 Flash is currently ranked first with 63.0%.
What does HiddenMath measure?
Google DeepMind's internal mathematical reasoning benchmark that introduces novel problems not encountered during model training to evaluate true mathematical reasoning capabilities rather than memorization
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
13 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.