What happened on the trip?
Guess what really happened on a two-day trip from its plan alone.
Run Nov 2, 2026, 5 items, $6.16 to run.
Method
Two travellers followed a plan for a weekend in the Baie de Somme, France, at the end of October 2026: trains, two taxis, a bird park, a night in Le Crotoy, a guided walk on the bay and a heritage train home. After the trip they wrote five questions about what actually happened and answered them before any model saw the questions.
Each model sees only the anonymized plan, including the packing list, and one question at a time, and gives a single best guess: what surprised them in town at 18:00, what disappointed them on the walk, which missing item they wished they had packed, what was inconvenient after the walk and how they solved it, and what the two taxi rides cost.
One call per model and question and no tools. The first run requested medium reasoning with a 4,096-token output cap including reasoning; the second asked each model for its highest supported effort (max, or xhigh or high where that is the top setting) with a 32,768-token cap. Provider meanings of each setting can differ. Only models answering all five questions are ranked.
Results
| # | Model | Accuracy |
|---|---|---|
| 1 | Claude Opus 5.5 (max reasoning · 5 questions) | 69.8% |
| 2 | GPT-6 Astra Pro (max reasoning · 5 questions) | 69.8% |
| 3 | GPT-6.1 Sol (max reasoning · 5 questions) | 69.8% |
| 4 | GPT-5.6 Sol (medium reasoning · 5 questions) | 69.8% |
| 5 | GPT-6.1 Pro Sol (max reasoning · 5 questions) | 66.9% |
| 6 | Claude Fable 5.1 (max reasoning · 5 questions) | 59.8% |
| 7 | GPT-6 Astra (max reasoning · 5 questions) | 59.8% |
| 8 | GPT-6.1 Sol (medium reasoning · 5 questions) | 59.8% |
| 9 | Muse Wonder (max reasoning · 5 questions) | 58.7% |
| 10 | Claude Opus 5.5 (medium reasoning · 5 questions) | 58.7% |
| 11 | Gemini 3.8 Flash (medium reasoning · 5 questions) | 58.7% |
| 12 | GPT-6 Astra (medium reasoning · 5 questions) | 58.7% |
| 13 | GPT-6.1 Terra (max reasoning · 5 questions) | 58.4% |
| 14 | Claude Sonnet 5.5 (max reasoning · 5 questions) | 57.3% |
| 15 | Gemini 3.6 Flash (medium reasoning · 5 questions) | 57.3% |
| 16 | Gemini 3.7 Flash (medium reasoning · 5 questions) | 57.3% |
| 17 | DeepSeek V4.1 Flash (1017) (max reasoning · 5 questions) | 55.9% |
| 18 | Claude Fable 5.1 (medium reasoning · 5 questions) | 54.4% |
| 19 | Gemini 4 Argon (high reasoning · 5 questions) | 49.8% |
| 20 | Gemini 3.5 Flash (medium reasoning · 5 questions) | 48.4% |
| 21 | GPT-5.6 Luna (medium reasoning · 5 questions) | 47.3% |
| 22 | Gemini 3.1 Pro (Preview) (medium reasoning · 5 questions) | 45.9% |
| 23 | Gemini 3 Flash (Preview) (medium reasoning · 5 questions) | 43.0% |
| 24 | GPT-6 Sol (medium reasoning · 5 questions) | 35.5% |
| 25 | Grok 4.7 (xhigh reasoning · 5 questions) | 31.5% |
| 26 | Kimi K3.1 Pro (max reasoning · 5 questions) | 29.8% |
| 27 | Qwen3.8 Max Prime (medium reasoning · 5 questions) | 23.0% |
| 28 | Grok 4.7 (medium reasoning · 5 questions) | 23.0% |
| 29 | GLM-5.3 Flash (medium reasoning · 5 questions) | 21.5% |
| 30 | Claude Sonnet 5.5 (medium reasoning · 5 questions) | 16.2% |
- Five questions, one answer per model and question. Q1-Q4 earn 1, ½ or 0 from a judge model (GPT-6.1 Sol, low reasoning) applying fixed rubrics to the travellers' own answers; the taxi total earns 1 minus its relative error, floored at 0. The score is the mean credit as a percentage.
- Two runs share the frozen plan, questions, key and judge: a first panel at medium reasoning and a panel of current leaderboard models at their highest reasoning effort. Each row is labelled with its effort.
- One trip and five questions: a single different answer moves a score by up to 20 points, so close scores do not establish an ordering.
In the main leaderboard
These results are collected for the main leaderboard alongside external benchmarks. New scores enter the rating through the regular audited data updates. Comparisons with that rating therefore share some of its inputs.
Interactive version: theaggregate.ai/our-benchmarks/trip-outcomes · How It Works · Data refreshed daily, snapshot 2026-11-05.