Pelican Arena
We show this table for reference; we do not rank on it.
Kaggle's Pelican Arena asks every model for the same thing, an animated SVG of a pelican riding a bicycle, and people vote for the better drawing.
Evaluation results published by Kaggle. How Kaggle Benchmarks reruns tests.
Arena rating on Pelican Arena v1 — Latest evaluation November 13, 2026
We mirror the published arena rating view for Pelican Arena v1. GPT-6 Astra has the highest reported value at 1819. A rating gap smaller than the two rows' intervals is not a reliable difference. We do not use these results to rank models overall.
GPT-6 Astra
Max Reasoning; 445 votes; ±63; Evaluated 2026-11-13
GPT-6.1 Sol
Max Reasoning; 445 votes; ±62; Evaluated 2026-11-13
GPT-6.1 Sol
Xhigh Reasoning; 448 votes; ±61; Evaluated 2026-11-13
60 source rowsCross-domainCurrentDisplay onlyUpdated Latest evaluation November 13, 2026
Arena rating table (60 source rows)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | GPT-6 AstraCreator not reportedMax Reasoning; 445 votes; ±63; Evaluated 2026-11-13 | 1819 | Not reported | Unknown |
| 2 | GPT-6.1 SolCreator not reportedMax Reasoning; 445 votes; ±62; Evaluated 2026-11-13 | 1787 | Not reported | Unknown |
| 3 | GPT-6.1 SolCreator not reportedXhigh Reasoning; 448 votes; ±61; Evaluated 2026-11-13 | 1785 | Not reported | Unknown |
| 4 | GPT-6 AstraCreator not reportedHigh Reasoning; 440 votes; ±61; Evaluated 2026-11-13 | 1785 | Not reported | Unknown |
| 5 | GPT-6 AstraCreator not reportedLow Reasoning; 455 votes; ±60; Evaluated 2026-11-13 | 1756 | Not reported | Unknown |
| 6 | Claude Opus 5.5Creator not reportedMax Reasoning; 342 votes; ±63; Evaluated 2026-11-13 | 1736 | Not reported | Unknown |
| 7 | Claude Sonnet 5.5Creator not reportedMax Reasoning; 353 votes; ±62; Evaluated 2026-11-13 | 1726 | Not reported | Unknown |
| 8 | Claude Opus 5.5Creator not reportedXhigh Reasoning; 442 votes; ±58; Evaluated 2026-11-13 | 1696 | Not reported | Unknown |
| 9 | GPT-6 SolCreator not reportedMax Reasoning; 441 votes; ±58; Evaluated 2026-11-13 | 1674 | Not reported | Unknown |
| 10 | Claude Fable 5.1Creator not reportedMax Reasoning; 439 votes; ±58; Evaluated 2026-11-13 | 1672 | Not reported | Unknown |
| 11 | Gemini 4 ArgonCreator not reportedHigh Reasoning; 436 votes; ±58; Evaluated 2026-11-13 | 1664 | Not reported | Unknown |
| 12 | GPT-5.6 SolCreator not reportedMax Reasoning; 443 votes; ±57; Evaluated 2026-11-13 | 1640 | Not reported | Unknown |
| 13 | Gemini 4 ArgonCreator not reportedLow Reasoning; 438 votes; ±57; Evaluated 2026-11-12 | 1633 | Not reported | Unknown |
| 14 | Gemini 4 CarbonCreator not reportedHigh Reasoning; 258 votes; ±64; Evaluated 2026-11-13 | 1600 | Not reported | Unknown |
| 15 | GLM-5.3-FlashCreator not reported440 votes; ±57; Evaluated 2026-11-13 | 1599 | Not reported | Unknown |
| 16 | Qwen 3.8 MaxCreator not reportedXhigh Reasoning; 442 votes; ±56; Evaluated 2026-11-13 | 1592 | Not reported | Unknown |
| 17 | GPT-5.6 SolCreator not reportedXhigh Reasoning; 442 votes; ±56; Evaluated 2026-11-13 | 1588 | Not reported | Unknown |
| 18 | Claude Opus 5Creator not reportedXhigh Reasoning; 440 votes; ±56; Evaluated 2026-11-13 | 1583 | Not reported | Unknown |
| 19 | Gemini 3.8 FlashCreator not reportedHigh Reasoning; 586 votes; ±54; Evaluated 2026-11-13 | 1570 | Not reported | Unknown |
| 20 | Gemini 3.8 FlashCreator not reportedMedium Reasoning; 593 votes; ±53; Evaluated 2026-11-13 | 1569 | Not reported | Unknown |
| 21 | GLM-5.3Creator not reportedMax Reasoning; 444 votes; ±56; Evaluated 2026-11-13 | 1568 | Not reported | Unknown |
| 22 | Claude Opus 5.5Creator not reportedHigh Reasoning; 443 votes; ±56; Evaluated 2026-11-13 | 1568 | Not reported | Unknown |
| 23 | Claude Opus 5Creator not reportedMax Reasoning; 437 votes; ±56; Evaluated 2026-11-12 | 1561 | Not reported | Unknown |
| 24 | Claude Opus 5.5Creator not reportedMedium Reasoning; 347 votes; ±59; Evaluated 2026-11-13 | 1552 | Not reported | Unknown |
| 25 | GPT-6 SolCreator not reportedHigh Reasoning; 455 votes; ±56; Evaluated 2026-11-13 | 1549 | Not reported | Unknown |
| 26 | GPT-6 LunaCreator not reportedMax Reasoning; 441 votes; ±56; Evaluated 2026-11-13 | 1536 | Not reported | Unknown |
| 27 | GPT-5.6 SolCreator not reportedMedium Reasoning; 443 votes; ±56; Evaluated 2026-11-13 | 1506 | Not reported | Unknown |
| 28 | Gemini 3.6 FlashCreator not reportedMedium Reasoning; 448 votes; ±56; Evaluated 2026-11-13 | 1494 | Not reported | Unknown |
| 29 | Claude Sonnet 5.5Creator not reportedHigh Reasoning; 435 votes; ±56; Evaluated 2026-11-13 | 1494 | Not reported | Unknown |
| 30 | Kimi K3Creator not reportedMax Reasoning; 436 votes; ±56; Evaluated 2026-11-13 | 1494 | Not reported | Unknown |
| 31 | Qwen 3.8 MaxCreator not reported450 votes; ±56; Evaluated 2026-11-13 | 1490 | Not reported | Unknown |
| 32 | Claude Fable 5.1Creator not reportedHigh Reasoning; 434 votes; ±56; Evaluated 2026-11-12 | 1489 | Not reported | Unknown |
| 33 | GLM-5.3Creator not reportedHigh Reasoning; 451 votes; ±56; Evaluated 2026-11-13 | 1473 | Not reported | Unknown |
| 34 | Grok 4.6Creator not reportedHigh Reasoning; 344 votes; ±59; Evaluated 2026-11-11 | 1468 | Not reported | Unknown |
| 35 | Kimi K3Creator not reportedMedium Reasoning; 440 votes; ±56; Evaluated 2026-11-13 | 1468 | Not reported | Unknown |
| 36 | Qwen 3.8 MaxCreator not reportedMedium Reasoning; 445 votes; ±56; Evaluated 2026-11-13 | 1467 | Not reported | Unknown |
| 37 | GPT-6 LunaCreator not reportedHigh Reasoning; 435 votes; ±56; Evaluated 2026-11-13 | 1465 | Not reported | Unknown |
| 38 | GPT-5.6 SolCreator not reportedHigh Reasoning; 443 votes; ±56; Evaluated 2026-11-13 | 1464 | Not reported | Unknown |
| 39 | Claude Opus 5Creator not reportedHigh Reasoning; 444 votes; ±56; Evaluated 2026-11-13 | 1457 | Not reported | Unknown |
| 40 | Gemini 3.6 FlashCreator not reportedHigh Reasoning; 439 votes; ±56; Evaluated 2026-11-13 | 1452 | Not reported | Unknown |
| 41 | GLM-5.3Creator not reportedMedium Reasoning; 459 votes; ±56; Evaluated 2026-11-13 | 1428 | Not reported | Unknown |
| 42 | Claude Fable 5.1Creator not reportedLow Reasoning; 440 votes; ±56; Evaluated 2026-11-13 | 1422 | Not reported | Unknown |
| 43 | GPT-5.6 LunaCreator not reportedMax Reasoning; 451 votes; ±56; Evaluated 2026-11-13 | 1420 | Not reported | Unknown |
| 44 | Grok 4.6Creator not reportedXhigh Reasoning; 343 votes; ±59; Evaluated 2026-11-12 | 1420 | Not reported | Unknown |
| 45 | Kimi K3Creator not reportedHigh Reasoning; 432 votes; ±57; Evaluated 2026-11-12 | 1417 | Not reported | Unknown |
| 46 | Qwen 3.8 MaxCreator not reportedLow Reasoning; 438 votes; ±56; Evaluated 2026-11-13 | 1400 | Not reported | Unknown |
| 47 | Claude Opus 5Creator not reportedMedium Reasoning; 444 votes; ±57; Evaluated 2026-11-13 | 1371 | Not reported | Unknown |
| 48 | Grok 4.5Creator not reportedHigh Reasoning; 443 votes; ±57; Evaluated 2026-11-13 | 1358 | Not reported | Unknown |
| 49 | GPT-5.6 TerraCreator not reportedGPT 5.6 Terra Max Reasoning; 437 votes; ±58; Evaluated 2026-11-13 | 1342 | Not reported | Unknown |
| 50 | GPT-5.6 LunaCreator not reportedXHigh Reasoning; 440 votes; ±57; Evaluated 2026-11-13 | 1341 | Not reported | Unknown |
| 51 | Claude Haiku 5.5Creator not reported442 votes; ±58; Evaluated 2026-11-13 | 1325 | Not reported | Unknown |
| 52 | Gemini 3.8 FlashCreator not reportedLow Reasoning; 95 votes; ±91; Evaluated 2026-11-13 | 1310 | Not reported | Unknown |
| 53 | Grok 4.7Creator not reportedXhigh Reasoning; 442 votes; ±58; Evaluated 2026-11-12 | 1304 | Not reported | Unknown |
| 54 | GPT-5.6 LunaCreator not reportedHigh Reasoning; 447 votes; ±59; Evaluated 2026-11-13 | 1279 | Not reported | Unknown |
| 55 | GPT-5.6 TerraCreator not reportedMedium Reasoning; 448 votes; ±59; Evaluated 2026-11-13 | 1267 | Not reported | Unknown |
| 56 | GPT-5.6 TerraCreator not reportedXHigh Reasoning; 436 votes; ±60; Evaluated 2026-11-13 | 1252 | Not reported | Unknown |
| 57 | GPT-5.6 TerraCreator not reportedHigh Reasoning; 445 votes; ±60; Evaluated 2026-11-13 | 1234 | Not reported | Unknown |
| 58 | Claude Sonnet 5Creator not reported435 votes; ±61; Evaluated 2026-11-13 | 1229 | Not reported | Unknown |
| 59 | GPT-5.6 LunaCreator not reportedMedium Reasoning; 440 votes; ±62; Evaluated 2026-11-13 | 1208 | Not reported | Unknown |
| 60 | Gemini 3.5 Flash-LiteCreator not reportedHigh Reasoning; 437 votes; ±63; Evaluated 2026-11-13 | 1186 | Not reported | Unknown |
How to read this leaderboard
Read the rating as a preference ranking for SVG drawings made from code among the models on this board, with its vote count and interval. It does not measure accuracy.
Operator receipt: 60 sourced rows are currently displayable on this page; the highest published score among these 60 source rows is GPT-6 Astra at 1819, 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 Pelican Arena
A drawing qualifies only if the SVG renders; qualifying drawings then go to human votes. It tests drawing with code: the model writes SVG markup, not an image. 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.
60 source rows are shown for Pelican 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. Pelican Arena is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Pelican Arena
Year
2026
Tasks
One prompt: an animated SVG of a pelican riding a bicycle, rated by human votes
Format
Relative rating fitted to human pairwise votes
Difficulty
Vector drawing and animation in code
A drawing qualifies only if the SVG renders; qualifying drawings then go to human votes. It tests drawing with code: the model writes SVG markup, not an image. 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
Pelican 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 Pelican Arena measure?
Kaggle's Pelican Arena asks every model for the same thing, an animated SVG of a pelican riding a bicycle, and people vote for the better drawing.
Which source row has the highest reported Pelican Arena value?
GPT-6 Astra has the highest reported Pelican Arena value at 1819. 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 Pelican Arena?
The Latest evaluation November 13, 2026 snapshot contains 60 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.