reasoning benchmark
SWE-bench Multilingual Leaderboard
SWE-bench Multilingual is a multilingual software engineering benchmark covering Java, TypeScript, JavaScript, Go, Rust, C, and C++.
Updated Sep 5, 2026
Models43
Model coverage43
MetricScore
EvidenceB
SWE-bench Multilingual Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAN | Score87.30% | Percentile100.00% | Participants43 | EvidenceC | Evaluated |
| Rank02 | ModelAN | Score84.40% | Percentile97.62% | Participants43 | EvidenceC | Evaluated |
| Rank03 | ModelTE | Score82.90% | Percentile95.24% | Participants43 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score81.00% | Percentile92.86% | Participants43 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score81.00% | Percentile90.48% | Participants43 | EvidenceC | Evaluated |
| Rank06 | ModelPO | Score78.50% | Percentile88.10% | Participants43 | EvidenceC | Evaluated |
| Rank07 | ModelAN | Score78.30% | Percentile85.71% | Participants43 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score78.30% | Percentile83.33% | Participants43 | EvidenceC | Evaluated |
| Rank09 | ModelAN | Score77.83% | Percentile80.95% | Participants43 | EvidenceC | Evaluated |
| Rank10 | ModelMA | Score76.70% | Percentile78.57% | Participants43 | EvidenceC | Evaluated |
| Rank11 | ModelMI | Score76.50% | Percentile76.19% | Participants43 | EvidenceC | Evaluated |
| Rank12 | ModelDE | Score76.20% | Percentile73.81% | Participants43 | EvidenceC | Evaluated |
| Rank13 | ModelTE | Score75.80% | Percentile71.43% | Participants43 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score75.80% | Percentile69.05% | Participants43 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score73.80% | Percentile66.67% | Participants43 | EvidenceC | Evaluated |
| Rank16 | ModelDE | Score73.30% | Percentile64.29% | Participants43 | EvidenceC | Evaluated |
| Rank17 | ModelMA | Score73.00% | Percentile61.90% | Participants43 | EvidenceC | Evaluated |
| Rank18 | ModelMI | Score72.50% | Percentile59.52% | Participants43 | EvidenceC | Evaluated |
| Rank19 | ModelXI | Score71.70% | Percentile57.14% | Participants43 | EvidenceC | Evaluated |
| Rank20 | ModelXI | Score71.70% | Percentile54.76% | Participants43 | EvidenceC | Evaluated |
| Rank21 | ModelAC | Score71.30% | Percentile52.38% | Participants43 | EvidenceC | Evaluated |
| Rank22 | ModelDE | Score70.20% | Percentile50.00% | Participants43 | EvidenceC | Evaluated |
| Rank23 | ModelDE | Score70.20% | Percentile47.62% | Participants43 | EvidenceC | Evaluated |
| Rank24 | ModelDE | Score70.20% | Percentile45.24% | Participants43 | EvidenceC | Evaluated |
| Rank25 | ModelAC | Score69.30% | Percentile42.86% | Participants43 | EvidenceC | Evaluated |
| Rank26 | ModelNV | Score67.70% | Percentile40.48% | Participants43 | EvidenceC | Evaluated |
| Rank27 | ModelAC | Score67.20% | Percentile38.10% | Participants43 | EvidenceC | Evaluated |
| Rank28 | ModelZA | Score66.70% | Percentile35.71% | Participants43 | EvidenceC | Evaluated |
| Rank29 | ModelMI | Score65.50% | Percentile33.33% | Participants43 | EvidenceC | Evaluated |
| Rank30 | ModelPO | Score63.10% | Percentile30.95% | Participants43 | EvidenceC | Evaluated |
SWE-bench Multilingual Highlights
The leading models and scores on this benchmark.
SWE-bench Multilingual Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for SWE-bench Multilingual
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is SWE-bench Multilingual?
What SWE-bench Multilingual measures and how its scores work.
SWE-bench Multilingual contains 1,632 instances annotated from 2,456 candidates by 68 expert annotators to evaluate Large Language Models across diverse software ecosystems beyond Python.
It measures issue resolving in software engineering using the Score metric, reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- SWE-bench Multilingual
- Modality
- text
- Primary category
- reasoning
- 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-bench Multilingual.
Which model scores highest on SWE-bench Multilingual?
Claude Mythos Preview is currently ranked first with 87.30%.
What are the top three models on SWE-bench Multilingual?
The current leaders are Claude Mythos Preview (87.30%), Claude Opus 4.8 (84.40%), and Hy4 preview (82.90%).
Which SWE-bench Multilingual model has the lowest official input price?
DeepSeek-V4-Flash-Max has the lowest matched official input price at $0.14 input / $0.28 output per 1M tokens.
Which models are fastest among SWE-bench Multilingual results?
The fastest matched records are Claude Opus 4.6 (123.15 tok/s via Anthropic), MiniMax M2.1 (100.00 tok/s via MiniMax), and DeepSeek-V3.2 (97.00 tok/s via DeepInfra).
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.