WildTrace: leaderboard
WildTrace (2026) asks multi-hop questions whose evidence is spread naturally across long incident reports and lesser-known literary works, supplied in full at up to about a million tokens. Answers are graded against criterion-level rubrics by three non-contestant LLM judges; the score is mean rubric credit in percent, a document longer than the model context window scoring zero, and higher is better.
Metric: Rubric credit (%; mean of three non-contestant LLM judges (Claude Sonnet 4.6, Qwen3.5-Plus, Gemini 2.5 Flash) grading answers against criterion-level rubrics, over 481 multi-hop questions on 214 long technical reports and literary works given in full without evidence spans, up to about 1M tokens; a document longer than the model context window is not sent and scores zero; higher is better). Source: arxiv.org. Saturation forecast: Around February 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 75.3 | #94 |
| 2 | Claude Opus 4.8 | 72.6 | #56 |
| 3 | GPT-5.5 | 71 | #49 |
| 4 | GLM-5.2 | 70.9 | #138 |
| 5 | Qwen 3.7 Max | 70.1 | #145 |
| 6 | Claude Opus 4.6 | 69.2 | #102 |
| 7 | Qwen 3.6 Plus | 66.7 | #222 |
| 8 | Qwen 3.7 Plus | 65.7 | #207 |
| 9 | GPT-5.4 | 62.2 | #123 |
| 10 | Gemini 2.5 Pro | 61.4 | #266 |
| 11 | GPT-4.1 | 60.4 | #495 |
| 12 | GPT-5.1 | 42.1 | #229 |
| 13 | Qwen 3 Max | 35.2 | #403 |
| 14 | Kimi K2.6 | 35.1 | #178 |
| 15 | DeepSeek V3.2 | 33 | #406 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=wildtrace · How It Works · Data refreshed daily, snapshot 2026-11-13.