ASCII Arena
We show this table for reference; we do not rank on it.
Kaggle's ASCII Arena asks models to draw everyday subjects as ASCII art, and people vote for the better drawing.
Evaluation results published by Kaggle. How Kaggle Benchmarks reruns tests.
Arena rating on ASCII Arena v1 — Latest evaluation November 13, 2026
We mirror the published arena rating view for ASCII Arena v1. Gemini 4 Carbon has the highest reported value at 2019. A rating gap smaller than the two rows' intervals is not a reliable difference. We do not use these results to rank models overall.
Gemini 4 Carbon
High Reasoning; 173 votes; ±127; Evaluated 2026-11-13
Gemini 4 Argon
High Reasoning; 411 votes; ±99; Evaluated 2026-11-11
GPT-6 Astra
High Reasoning; 427 votes; ±86; Evaluated 2026-11-13
18 source rowsCross-domainCurrentDisplay onlyUpdated Latest evaluation November 13, 2026
Arena rating table (18 source rows)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Gemini 4 CarbonCreator not reportedHigh Reasoning; 173 votes; ±127; Evaluated 2026-11-13 | 2019 | Not reported | Unknown |
| 2 | Gemini 4 ArgonCreator not reportedHigh Reasoning; 411 votes; ±99; Evaluated 2026-11-11 | 1983 | Not reported | Unknown |
| 3 | GPT-6 AstraCreator not reportedHigh Reasoning; 427 votes; ±86; Evaluated 2026-11-13 | 1730 | Not reported | Unknown |
| 4 | GPT-6.1 SolCreator not reportedHigh Reasoning; 437 votes; ±85; Evaluated 2026-11-13 | 1686 | Not reported | Unknown |
| 5 | Gemini 3.8 FlashCreator not reportedHigh Reasoning; 405 votes; ±85; Evaluated 2026-11-11 | 1672 | Not reported | Unknown |
| 6 | Claude Opus 5.5Creator not reportedHigh Reasoning; 445 votes; ±84; Evaluated 2026-11-13 | 1651 | Not reported | Unknown |
| 7 | Gemini 3.1 Pro PreviewCreator not reportedHigh Reasoning; 454 votes; ±83; Evaluated 2026-11-13 | 1573 | Not reported | Unknown |
| 8 | Claude Sonnet 5.5Creator not reportedHigh Reasoning; 419 votes; ±83; Evaluated 2026-11-13 | 1541 | Not reported | Unknown |
| 9 | Claude Opus 5Creator not reportedHigh Reasoning; 427 votes; ±83; Evaluated 2026-11-12 | 1506 | Not reported | Unknown |
| 10 | GPT-6 SolCreator not reportedHigh Reasoning; 445 votes; ±82; Evaluated 2026-11-11 | 1497 | Not reported | Unknown |
| 11 | Claude Fable 5.1Creator not reportedHigh Reasoning; 424 votes; ±83; Evaluated 2026-11-13 | 1484 | Not reported | Unknown |
| 12 | Claude Haiku 5.5Creator not reportedHigh Reasoning; 409 votes; ±84; Evaluated 2026-11-12 | 1354 | Not reported | Unknown |
| 13 | Kimi K3Creator not reportedHigh Reasoning; 410 votes; ±84; Evaluated 2026-11-13 | 1303 | Not reported | Unknown |
| 14 | GPT-5.6 TerraCreator not reportedHigh Reasoning; 445 votes; ±84; Evaluated 2026-11-12 | 1294 | Not reported | Unknown |
| 15 | Qwen 3.8 MaxCreator not reportedXhigh Reasoning; 429 votes; ±84; Evaluated 2026-11-11 | 1282 | Not reported | Unknown |
| 16 | GPT-6 LunaCreator not reportedHigh Reasoning; 447 votes; ±83; Evaluated 2026-11-13 | 1279 | Not reported | Unknown |
| 17 | Gemini 3.5 Flash-LiteCreator not reportedHigh Reasoning; 423 votes; ±84; Evaluated 2026-11-13 | 1271 | Not reported | Unknown |
| 18 | GPT-5.6 LunaCreator not reportedHigh Reasoning; 446 votes; ±84; Evaluated 2026-11-13 | 1262 | Not reported | Unknown |
How to read this leaderboard
Read the rating as a preference ranking for ASCII art among the models on this board, with its vote count and interval. It does not measure accuracy.
Operator receipt: 18 sourced rows are currently displayable on this page; the highest published score among these 18 source rows is Gemini 4 Carbon at 2019, which does not establish a market leader.
Honest limit: 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.
How to read ASCII Arena
Subjects come from Google's Quick, Draw! dataset. It tests visual layout using plain text characters only. 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.
Snapshot
A rating gap smaller than the two rows' intervals is not a reliable difference. The table retains each source configuration and its reported value.
18 source rows are shown for ASCII Arena. The benchmark falls in the Cross-domain category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. ASCII Arena is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About ASCII Arena
Year
2026
Tasks
ASCII drawings of subjects from Google's Quick, Draw! dataset, rated by human votes
Format
Relative rating fitted to human pairwise votes
Difficulty
Text-character drawing
Subjects come from Google's Quick, Draw! dataset. It tests visual layout using plain text characters only. 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.
Freshness and provenance
Version
ASCII Arena 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does ASCII Arena measure?
Kaggle's ASCII Arena asks models to draw everyday subjects as ASCII art, and people vote for the better drawing.
Which source row has the highest reported ASCII Arena value?
Gemini 4 Carbon has the highest reported ASCII Arena value at 2019. A rating gap smaller than the two rows' intervals is not a reliable difference. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on ASCII Arena?
The Latest evaluation November 13, 2026 snapshot contains 18 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.