reasoning benchmark
Agents' Last Exam
Agents' Last Exam is a challenging benchmark for AI agents on hard, long-horizon tasks that test sustained reasoning, planning, and tool use, reported with and without tool access.
Updated Aug 7, 2026
Published models5
Registry coverage5
MetricScore
EvidenceB
Agents' Last Exam leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | OP | 52.7% | 100.0% | 5 | C | |
| 2 | OP | 50.4% | 75.0% | 5 | C | |
| 3 | OP | 50.3% | 50.0% | 5 | C | |
| 4 | BY | 41.4% | 25.0% | 5 | C | |
| 5 | DE | 25.2% | 0.0% | 5 | C |
Score distribution
Top published rows on the benchmark's original scale.
Agents' Last Exam highlights
The top published results on this benchmark's own scale.
What is Agents' Last Exam?
Definition and scoring fields from the benchmark registry.
Agents' Last Exam is a challenging benchmark for AI agents on hard, long-horizon tasks that test sustained reasoning, planning, and tool use, reported with and without tool access.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- Agents' Last Exam
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- agents-last-exam|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about Agents' Last Exam.
Which model scores highest on Agents' Last Exam?
GPT-5.6 Sol is currently ranked first with 52.7%.
What does Agents' Last Exam measure?
Agents' Last Exam is a challenging benchmark for AI agents on hard, long-horizon tasks that test sustained reasoning, planning, and tool use, reported with and without tool access.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
5 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.