agents benchmark
FrontierSWE
FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. Performance is reported as a dominance score, where higher is better.
Updated Aug 7, 2026
Published models15
Registry coverage15
MetricScore
EvidenceB
FrontierSWE leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AN | 90.0% | 100.0% | 15 | B | |
| 2 | MA | 81.2% | 92.9% | 15 | C | |
| 3 | AN | 75.0% | 85.7% | 15 | B | |
| 4 | ZA | 74.0% | 78.6% | 15 | B | |
| 5 | AC | 73.5% | 71.4% | 15 | C | |
| 6 | OP | 73.0% | 64.3% | 15 | B | |
| 7 | AN | 63.0% | 57.1% | 15 | B | |
| 8 | AN | 56.0% | 50.0% | 15 | B | |
| 9 | OP | 54.0% | 42.9% | 15 | B | |
| 10 | GO | 40.0% | 35.7% | 15 | B | |
| 11 | ZA | 31.0% | 28.6% | 15 | B | |
| 12 | DE | 29.0% | 21.4% | 15 | B | |
| 13 | MA | 27.0% | 14.3% | 15 | B | |
| 14 | MA | 26.0% | 7.1% | 15 | B | |
| 15 | AC | 22.0% | 0.0% | 15 | B |
Score distribution
Top published rows on the benchmark's original scale.
FrontierSWE highlights
The top published results on this benchmark's own scale.
What is FrontierSWE?
Definition and scoring fields from the benchmark registry.
FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. Performance is reported as a dominance score, where higher is better.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- FrontierSWE
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- frontierswe|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about FrontierSWE.
Which model scores highest on FrontierSWE?
Claude Fable 5 is currently ranked first with 90.0%.
What does FrontierSWE measure?
FrontierSWE measures whether an agent can complete open-ended technical projects at the scale of hours to tens of hours, spanning systems optimization, large-scale code construction, and applied ML research. Performance is reported as a dominance score, where higher is better.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
15 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.