# ELI5 Arena

> Kaggle's ELI5 Arena asks models to explain 100 topics the way you would to a five-year-old, and people vote for the better explanation.

Canonical page: https://benchlm.ai/benchmarks/eli5-arena

- Category: [external](/external)
- Last updated: Latest evaluation November 12, 2026

Evaluation results published by [Kaggle](https://www.kaggle.com/benchmarks/kaggle/eli5-arena). undefined

A response qualifies if it is a plain-text answer of 20 to 300 words. Kaggle shows two anonymous outputs for the same prompt and a person picks the better one. The Arena column is the relative rating Kaggle fits to those votes, published with an interval; it is not a percentage.

The rating moves as votes arrive, and Kaggle publishes no confidence level for its interval. Rows with few votes have wide intervals, so neighbouring models are often not separable. Voters are self-selected Kaggle visitors. The board declares no license; we keep published aggregates with attribution and no task or output contents.

A rating gap smaller than the two rows' intervals is not a reliable difference.

- [Kaggle arena leaderboard](https://www.kaggle.com/benchmarks/kaggle/eli5-arena)
- [Published leaderboard data](https://www.kaggle.com/api/v1/benchmarks/kaggle/eli5-arena/leaderboard)

## About ELI5 Arena

- Year: 2026
- Tasks: 100 "explain it like I'm five" prompts in 10 categories, rated by human votes
- Format: Relative rating fitted to human pairwise votes
- Difficulty: Plain-language explanation

A response qualifies if it is a plain-text answer of 20 to 300 words. Kaggle shows two anonymous outputs for the same prompt and a person picks the better one. The Arena column is the relative rating Kaggle fits to those votes, published with an interval; it is not a percentage. These results are display-only and do not enter overall or category rankings.

ELI5 Arena is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (19 source rows)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Gemini 4 Argon](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | High Reasoning; 7 votes; ±239; Evaluated 2026-11-12 | — | 1691 |
| 2 | [GPT-5.6 Sol](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 5 votes; ±259; Evaluated 2026-11-10 | — | 1689 |
| 3 | [Qwen 3.8 Max](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Xhigh Reasoning; 10 votes; ±209; Evaluated 2026-11-12 | — | 1656 |
| 4 | [GLM-5.3](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 5 votes; ±245; Evaluated 2026-11-12 | — | 1596 |
| 5 | [GPT-5.6 Terra](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | High Reasoning; 13 votes; ±187; Evaluated 2026-11-12 | — | 1591 |
| 6 | [Gemini 3.8 Flash](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | High Reasoning; 5 votes; ±240; Evaluated 2026-11-12 | — | 1583 |
| 7 | [Claude Fable 5.1](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 8 votes; ±213; Evaluated 2026-11-12 | — | 1567 |
| 8 | [Gemini 3.1 Pro Preview](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | High Reasoning; 9 votes; ±206; Evaluated 2026-11-12 | — | 1468 |
| 9 | [GPT-6 Astra](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 15 votes; ±177; Evaluated 2026-11-10 | — | 1463 |
| 10 | [GPT-5 Pro](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | High Reasoning; 5 votes; ±241; Evaluated 2026-11-10 | — | 1463 |
| 11 | [GPT-5.5](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) |  GPT-5.5 High Reasoning; 6 votes; ±228; Evaluated 2026-11-12 | — | 1450 |
| 12 | [Claude Opus 5.5](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 7 votes; ±223; Evaluated 2026-11-12 | — | 1444 |
| 13 | [Claude Opus 5](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 5 votes; ±240; Evaluated 2026-11-12 | — | 1436 |
| 14 | [GPT-6 Luna](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 9 votes; ±208; Evaluated 2026-11-12 | — | 1422 |
| 15 | [Claude Sonnet 5](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | High Reasoning; 5 votes; ±247; Evaluated 2026-11-12 | — | 1416 |
| 16 | [Claude Sonnet 5.5](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 9 votes; ±211; Evaluated 2026-11-12 | — | 1407 |
| 17 | [GPT-5.6 Terra](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | GPT 5.6 Terra Max Reasoning; 11 votes; ±204; Evaluated 2026-11-10 | — | 1375 |
| 18 | [Grok 4.7](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Xhigh Reasoning; 7 votes; ±236; Evaluated 2026-11-10 | — | 1371 |
| 19 | [GPT-6.1 Sol](https://www.kaggle.com/benchmarks/kaggle/eli5-arena) | Max Reasoning; 6 votes; ±238; Evaluated 2026-11-10 | — | 1344 |

## FAQ

### What does ELI5 Arena measure?

Kaggle's ELI5 Arena asks models to explain 100 topics the way you would to a five-year-old, and people vote for the better explanation.

### Which source row has the highest reported ELI5 Arena value?

Gemini 4 Argon has the highest reported ELI5 Arena value at 1691. A rating gap smaller than the two rows' intervals is not a reliable difference.

### How many models are evaluated on ELI5 Arena?

The Latest evaluation November 12, 2026 contains 19 AI models.

### Does ELI5 Arena affect BenchLM's overall score?

Not directly. ELI5 Arena is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.