agents benchmark
SWE-Marathon Leaderboard
SWE-Marathon is an ultra-long-horizon software engineering benchmark covering tasks such as building compilers, optimizing kernels, and developing production-grade services.
Updated Oct 23, 2026
Models10
Model coverage10
MetricScore
EvidenceB
SWE-Marathon Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelME | Score57.50% | Percentile100.00% | Participants10 | EvidenceC | Evaluated |
| Rank02 | ModelME | Score57.50% | Percentile88.89% | Participants10 | EvidenceC | Evaluated |
| Rank03 | ModelME | Score53.10% | Percentile77.78% | Participants10 | EvidenceC | Evaluated |
| Rank04 | ModelME | Score53.10% | Percentile66.67% | Participants10 | EvidenceC | Evaluated |
| Rank05 | ModelXA | Score46.00% | Percentile55.56% | Participants10 | EvidenceC | Evaluated |
| Rank06 | ModelZA | Score42.50% | Percentile44.44% | Participants10 | EvidenceC | Evaluated |
| Rank07 | ModelMA | Score42.00% | Percentile33.33% | Participants10 | EvidenceC | Evaluated |
| Rank08 | ModelTE | Score31.90% | Percentile22.22% | Participants10 | EvidenceC | Evaluated |
| Rank09 | ModelXA | Score29.00% | Percentile11.11% | Participants10 | EvidenceC | Evaluated |
| Rank10 | ModelZA | Score13.00% | Percentile0.00% | Participants10 | EvidenceC | Evaluated |
SWE-Marathon Highlights
The leading models and scores on this benchmark.
SWE-Marathon Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for SWE-Marathon
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is SWE-Marathon?
What SWE-Marathon measures and how its scores work.
SWE-Marathon is a software engineering benchmark for evaluating agents on tasks such as building compilers, optimizing kernels, and developing production-grade services.
It measures whether agents can sustain quality across extremely long engineering trajectories, using the Score metric with a ratio unit.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- SWE-Marathon
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about SWE-Marathon.
Which model scores highest on SWE-Marathon?
Muse Wonder Full is currently ranked first with 57.50%.
What are the top three models on SWE-Marathon?
The current leaders are Muse Wonder Full (57.50%), Muse Wonder Full (57.50%), and Muse Wonder (53.10%).
Which SWE-Marathon model has the lowest official input price?
GLM-5.3 has the lowest matched official input price at $1.2 input / $4.2 output per 1M tokens.
Which models are fastest among SWE-Marathon results?
The fastest matched records are Grok 4.5 (80.00 tok/s via xAI).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.