reasoning benchmark
SWE-bench Verified (Agentless)
A human-validated subset of SWE-bench that evaluates language models' ability to resolve real-world GitHub issues using an agentless approach. The benchmark tests models on software engineering problems requiring understanding and coordinating changes across multiple functions, classes, and files simultaneously.
Updated Aug 7, 2026
Published models2
Registry coverage2
MetricScore
EvidenceB
SWE-bench Verified (Agentless) leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MA | 51.8% | 100.0% | 2 | C | |
| 2 | XI | 35.7% | 0.0% | 2 | C |
Score distribution
Top published rows on the benchmark's original scale.
SWE-bench Verified (Agentless) highlights
The top published results on this benchmark's own scale.
What is SWE-bench Verified (Agentless)?
Definition and scoring fields from the benchmark registry.
A human-validated subset of SWE-bench that evaluates language models' ability to resolve real-world GitHub issues using an agentless approach. The benchmark tests models on software engineering problems requiring understanding and coordinating changes across multiple functions, classes, and files simultaneously.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- SWE-bench Verified (Agentless)
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- swe-bench-verified-(agentless)|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about SWE-bench Verified (Agentless).
Which model scores highest on SWE-bench Verified (Agentless)?
Kimi K2 Instruct is currently ranked first with 51.8%.
What does SWE-bench Verified (Agentless) measure?
A human-validated subset of SWE-bench that evaluates language models' ability to resolve real-world GitHub issues using an agentless approach. The benchmark tests models on software engineering problems requiring understanding and coordinating changes across multiple functions, classes, and files simultaneously.
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.