RandomBench
Models repeatedly choose an integer from 0 to 99. The benchmark measures how evenly they use the range and how often they return valid numbers.
Kaggle publishes scores for 37 distinct model versions.
- Made by
- Benjamin Brumfield
- Released
- November 2026
- Kind
- Abstract puzzles and logic
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-6.2 Astra #1 on the capability index
- Last model added
- November 2026
- Scores refreshed
- November 20th, 2026
Grok 4.6 leads with 0.42
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
- 0.42 Grok 4.6
- 0.39 GPT-5.4
- 0.37
Gemini 4 Neon
- 0.35 Claude Opus 5.5
- 0.32 Grok 4.7
- 0.30 GPT-6.2 Astra
- 0.30 GPT-6.2 Sol
- 0.30
Gemini 3.8 Flash
- 0.30
Gemini 3.1 Pro
- 0.29 GPT-5.6 Terra
- 0.29 gpt-oss-120b
- 0.26
Gemini 3.6 Flash
- 0.26 GPT-5.4 Mini
- 0.24 GPT-5.4 Nano
- 0.23 GPT-6 Astra
- 0.23 GPT-6.1 Sol
- 0.23 Claude Opus 4.5
- 0.22 Claude Opus 4.7
- 0.21
Gemini 3 Flash
- 0.21
Gemma 4 26B A4B
- 0.20 Claude Haiku 5.5
- 0.20
Gemini 4 Argon
- 0.19
Gemma 4 31B IT
- 0.19 Claude Opus 4.8
- 0.17 Claude Opus 5
- 0.16 GPT-5.5
- 0.15 Claude Sonnet 4.6
- 0.15 GPT-6 Sol
- 0.14 GPT-5.6 Luna
- 0.13 Claude Haiku 4.5
- 0.12 Claude Sonnet 5
- 0.12 GPT-6 Luna
- 0.11 GPT-5.6 Sol
- 0.11 Claude Sonnet 5.5
- 0.11
Gemini 3.5 Flash-Lite
- 0.10 Claude Opus 4.6
- 0.10
Gemini 3.1 Flash-Lite
- 1 Grok 4.6 0.42
- 2 GPT-5.4 0.39
- 3
Gemini 4 Neon 0.37
- 4 Claude Opus 5.5 0.35
- 5 Grok 4.7 0.32
- 6 GPT-6.2 Astra 0.30
- 7 GPT-6.2 Sol 0.30
- 8
Gemini 3.8 Flash 0.30
- 9
Gemini 3.1 Pro 0.30
- 10 GPT-5.6 Terra 0.29
- 11 gpt-oss-120b 0.29
- 12
Gemini 3.6 Flash 0.26
- 13 GPT-5.4 Mini 0.26
- 14 GPT-5.4 Nano 0.24
- 15 GPT-6 Astra 0.23
- 16 GPT-6.1 Sol 0.23
- 17 Claude Opus 4.5 0.23
- 18 Claude Opus 4.7 0.22
- 19
Gemini 3 Flash 0.21
- 20
Gemma 4 26B A4B 0.21
- 21 Claude Haiku 5.5 0.20
- 22
Gemini 4 Argon 0.20
- 23
Gemma 4 31B IT 0.19
- 24 Claude Opus 4.8 0.19
- 25 Claude Opus 5 0.17
- 26 GPT-5.5 0.16
- 27 Claude Sonnet 4.6 0.15
- 28 GPT-6 Sol 0.15
- 29 GPT-5.6 Luna 0.14
- 30 Claude Haiku 4.5 0.13
- 31 Claude Sonnet 5 0.12
- 32 GPT-6 Luna 0.12
- 33 GPT-5.6 Sol 0.11
- 34 Claude Sonnet 5.5 0.11
- 35
Gemini 3.5 Flash-Lite 0.11
- 36 Claude Opus 4.6 0.10
- 37
Gemini 3.1 Flash-Lite 0.10
Source: Official RandomBench leaderboard
| # | Model | Score ± | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | Grok 4.6 | 0.42 ±0 | Aug 2026 | Nov 2026 | Official leaderboard | |
| 2 | GPT-5.4 | 0.39 ±0 | Mar 2026 | Nov 2026 | Official leaderboard | |
| 3 | | 0.37 ±0 | Nov 2026 | Nov 2026 | Official leaderboard | |
| 4 | Claude Opus 5.5 | 0.35 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 5 | Grok 4.7 | 0.32 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 6 | GPT-6.2 Astra | 0.30 ±0 | Nov 2026 | Nov 2026 | Official leaderboard | |
| 7 | GPT-6.2 Sol | 0.30 ±0 | Nov 2026 | Nov 2026 | Official leaderboard | |
| 8 | | 0.30 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 9 | | 0.30 ±0 | Feb 2026 | Nov 2026 | Official leaderboard | |
| 10 | GPT-5.6 Terra | 0.29 ±0 | Jul 2026 | Nov 2026 | Official leaderboard | |
| 11 | gpt-oss-120b | 0.29 ±0 | Aug 2025 | Nov 2026 | Official leaderboard | |
| 12 | | 0.26 ±0 | Jul 2026 | Nov 2026 | Official leaderboard | |
| 13 | GPT-5.4 Mini | 0.26 ±0 | Mar 2026 | Nov 2026 | Official leaderboard | |
| 14 | GPT-5.4 Nano | 0.24 ±0 | Mar 2026 | Nov 2026 | Official leaderboard | |
| 15 | GPT-6 Astra | 0.23 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 16 | GPT-6.1 Sol | 0.23 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 17 | Claude Opus 4.5 | 0.23 ±0 | Nov 2025 | Nov 2026 | Official leaderboard | |
| 18 | Claude Opus 4.7 | 0.22 ±0 | Apr 2026 | Nov 2026 | Official leaderboard | |
| 19 | | 0.21 ±0 | Dec 2025 | Nov 2026 | Official leaderboard | |
| 20 | | 0.21 ±0 | Apr 2026 | Nov 2026 | Official leaderboard | |
| 21 | Claude Haiku 5.5 | 0.20 ±0 | Oct 2026 | Nov 2026 | Official leaderboard | |
| 22 | | 0.20 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 23 | | 0.19 ±0 | Apr 2026 | Nov 2026 | Official leaderboard | |
| 24 | Claude Opus 4.8 | 0.19 ±0 | May 2026 | Nov 2026 | Official leaderboard | |
| 25 | Claude Opus 5 | 0.17 ±0 | Jul 2026 | Nov 2026 | Official leaderboard | |
| 26 | GPT-5.5 | 0.16 ±0 | Apr 2026 | Nov 2026 | Official leaderboard | |
| 27 | Claude Sonnet 4.6 | 0.15 ±0 | Feb 2026 | Nov 2026 | Official leaderboard | |
| 28 | GPT-6 Sol | 0.15 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 29 | GPT-5.6 Luna | 0.14 ±0 | Jul 2026 | Nov 2026 | Official leaderboard | |
| 30 | Claude Haiku 4.5 | 0.13 ±0 | Oct 2025 | Nov 2026 | Official leaderboard | |
| 31 | Claude Sonnet 5 | 0.12 ±0 | Jun 2026 | Nov 2026 | Official leaderboard | |
| 32 | GPT-6 Luna | 0.12 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 33 | GPT-5.6 Sol | 0.11 ±0 | Jul 2026 | Nov 2026 | Official leaderboard | |
| 34 | Claude Sonnet 5.5 | 0.11 ±0 | Sep 2026 | Nov 2026 | Official leaderboard | |
| 35 | | 0.11 ±0 | Jul 2026 | Nov 2026 | Official leaderboard | |
| 36 | Claude Opus 4.6 | 0.10 | Feb 2026 | Nov 2026 | Official leaderboard | |
| 37 | | 0.10 | Mar 2026 | Nov 2026 | Official leaderboard |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official RandomBench leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official RandomBench leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Science and reasoning
- GPQA Diamond In use 198 graduate-level 'Google-proof' multiple-choice questions in biology, chemistry and physics.
Gemini 4 Carbon 97%
- Terminal-Bench-Science In use 70 scientific research-workflow tasks across five science areas, run in Harbor containers. Hidden tests give the resolution rate. GPT-6 Astra 68.1%
- BioMysteryBench In use Bioinformatics problems on anonymized real datasets that require real analysis with tools. Rubric-graded, with human-solvable and human-difficult splits. Claude Fable 5.2 83.7%
- FrontierScience In use Models solve physics, chemistry, and biology problems in an Olympiad track graded by short answer and a Research track graded by 10-point rubrics.
Intern-S2-397B 87%
- HiPhO In use Models solve theory problems from 2024 and 2025 physics olympiad exams, some with diagrams. Official marking schemes grade answers and steps, mapped to medal levels.
Gemini 3 Pro 78.3%
- LABBench2 In use 11 biology research sub-tasks (literature, figures, tables, patents, trials and more) with tool access. Macro-averaged accuracy. GPT-5.2 60.9%
Data audit
All 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (37 entries).
- Pass Models scored 37 models we track.
- Pass Newest score The newest is from November 2026, for Gemini 4 Neon.
- Pass Frontier models 6 of the top 10 have a score, led by GPT-6.2 Astra (#1).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on kaggle.com: 37 rows for models we track at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.