reasoning benchmark
Agent Startup Bench
Agent Startup Bench measures AI agents on high-economic-value, startup-style tasks that require autonomous planning and execution to deliver practical, verifiable results.
Updated Aug 7, 2026
Published models2
Registry coverage2
MetricScore
EvidenceB
Agent Startup Bench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | BY | 68.8% | 100.0% | 2 | C | |
| 2 | BY | 54.0% | 0.0% | 2 | C |
Score distribution
Top published rows on the benchmark's original scale.
Agent Startup Bench highlights
The top published results on this benchmark's own scale.
What is Agent Startup Bench?
Definition and scoring fields from the benchmark registry.
Agent Startup Bench measures AI agents on high-economic-value, startup-style tasks that require autonomous planning and execution to deliver practical, verifiable results.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- Agent Startup Bench
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- agent-startup-bench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about Agent Startup Bench.
Which model scores highest on Agent Startup Bench?
Seed 2.1 Pro is currently ranked first with 68.8%.
What does Agent Startup Bench measure?
Agent Startup Bench measures AI agents on high-economic-value, startup-style tasks that require autonomous planning and execution to deliver practical, verifiable results.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
2 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.