SWE Multilingual
A multilingual software-engineering benchmark for real-world code issue resolution across multiple programming languages.
Top models on SWE Multilingual — October 29, 2026
As of October 29, 2026, Claude Opus 5.5 leads the SWE Multilingual leaderboard with 93.9% , followed by Claude Sonnet 5.5 (90.3%) and Claude Opus 5 (89.5%).
Claude Opus 5.5
Anthropic
Claude Sonnet 5.5
Anthropic
Claude Opus 5
Anthropic
60 modelsCoding5% of Coding reference weightCurrentUpdated October 29, 2026
Leaderboard (60 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Claude Opus 5.5Anthropic | 93.9% | Not reported | Closed |
| 2 | Claude Sonnet 5.5Anthropic | 90.3% | Not reported | Closed |
| 3 | Claude Opus 5Anthropic | 89.5% | Not reported | Closed |
| 4 | Claude Fable 5.1Anthropic | 89.1% | Not reported | Closed |
| 5 | Ling 3.1 FlashInclusionAI | 88.5% | Not reported | Open |
| 6 | Claude Opus 4.8Anthropic | 84.4% | Not reported | Closed |
| 7 | Hy4 previewTencent | 82.9% | Not reported | Open |
| 8 | Qwen3.8-Flash-NextAlibaba | 81% | Not reported | Open |
| 9 | Qwen3.8-Omni-FlashAlibaba | 80.5% | Not reported | Closed |
| 10 | Composer 2.5Cursor | 79.8% | Not reported | Closed |
| 11 | Ornith-1.5-397BOrnith AI | 79.6% | Not reported | Open |
| 12 | Ornith-1.0-397BDeepReinforce AI | 78.9% | Not reported | Open |
| 13 | Laguna S 2.1Poolside | 78.5% | Not reported | Open |
| 14 | Qwen3.7 MaxAlibaba | 78.3% | Not reported | Closed |
| 15 | Claude Sonnet 5Anthropic | 78.3% | Not reported | Closed |
| 16 | Beam 1.0 501B-A23BReflection AI | 78.3% | Not reported | Open |
| 17 | Beam 1.0 501B-A23B FP8Reflection AI | 78.1% | Not reported | Open |
| 18 | Grok 4.5xAI | 78% | Not reported | Closed |
| 19 | SWE-1.7Cognition | 77.8% | Not reported | Closed |
| 20 | Claude Opus 4.5Anthropic | 77.5% | Not reported | Closed |
| 21 | Beam 1.0 501B-A23B NVFP4Reflection AI | 77.5% | Not reported | Open |
| 22 | Kimi K2.6Moonshot AI | 76.7% | Not reported | Open |
| 23 | MiniMax M2.7MiniMax | 76.5% | Not reported | Open |
| 24 | DeepSeek V4 Pro 0813DeepSeek | 76.2% | Not reported | Open |
| 25 | Qwen3.7 PlusAlibaba | 75.8% | Not reported | Closed |
| 26 | dots3-note PreviewDots Studio | 75.7% | Not reported | Open |
| 27 | DeepSeek V4 Pro (High)DeepSeek | 74.1% | Not reported | Open |
| 28 | Qwen3.6 PlusAlibaba | 73.8% | Not reported | Closed |
| 29 | Composer 2Cursor | 73.7% | Not reported | Closed |
| 30 | GLM-5Z.AI | 73.3% | Not reported | Open |
| 31 | DeepSeek V4 Flash 0731DeepSeek | 73.3% | Not reported | Open |
| 32 | Kimi K2.5Moonshot AI | 73% | Not reported | Open |
| 33 | Sage-A1.5InternScience | 72.7% | Not reported | Open |
| 34 | Ling 3.0 FlashInclusionAI | 72.4% | Not reported | Open |
| 35 | MiMo-V2-FlashXiaomi | 71.7% | Not reported | Open |
| 36 | Ornith-1.5-35B-A3BOrnith AI | 71.4% | Not reported | Open |
| 37 | Qwen3.6-27BAlibaba | 71.3% | Not reported | Open |
| 38 | DeepSeek V4 Flash (High)DeepSeek | 70.2% | Not reported | Open |
| 39 | DeepSeek V4 ProDeepSeek | 69.8% | Not reported | Open |
| 40 | DeepSeek V4 FlashDeepSeek | 69.7% | Not reported | Open |
| 41 | Qwen3.5 397BAlibaba | 69.3% | Not reported | Open |
| 42 | Ornith-1.0-35BDeepReinforce AI | 69.3% | Not reported | Open |
| 43 | Nemotron 3.5 Super VL 120B A12BNVIDIA | 67.7% | Not reported | Open |
| 44 | Nemotron 3 UltraNVIDIA | 67.7% | Not reported | Open |
| 45 | Qwen3.6-35B-A3BAlibaba | 67.2% | Not reported | Open |
| 46 | Nemotron 3.5 Super VL 120B A12B NVFP4NVIDIA | 67% | Not reported | Open |
| 47 | Spark-X2.6-4BSparkLLM | 64.7% | Not reported | Open |
| 48 | Laguna M.1Poolside | 63.1% | Not reported | Closed |
| 49 | Laguna XS 2.1Poolside | 63.1% | Not reported | Open |
| 50 | LongCat-Flash-Lite-SparseMeituan | 59.3% | Not reported | Open |
| 51 | Laguna XS.2Poolside | 57.7% | Not reported | Open |
| 52 | Qwen3-Coder 480B-A35BAlibaba | 54.7% | Not reported | Open |
| 53 | Ornith-1.5-9BOrnith AI | 54.4% | Not reported | Open |
| 54 | Ornith-1.0-9BDeepReinforce AI | 52% | Not reported | Open |
| 55 | Kimi K2Moonshot AI | 47.3% | Not reported | Closed |
| 56 | Granite 4.2 30BIBM | 41.9% | Not reported | Open |
| 57 | Nemotron 3.5 Lightning 30B A3B NVFP4NVIDIA | 36.5% | Not reported | Open |
| 58 | Granite 4.2 8BIBM | 30.8% | Not reported | Open |
| 59 | LLaDA2.2-flashInclusionAI | 25% | Not reported | Open |
| 60 | LFM3-1.7BLiquidAI | 23% | Not reported | Open |
According to BenchLM.ai, Claude Opus 5.5 leads the SWE Multilingual benchmark with a score of 93.9%, followed by Claude Sonnet 5.5 (90.3%) and Claude Opus 5 (89.5%). The scores show moderate spread, with meaningful differences between the top tier and mid-tier models.
60 models have been evaluated on SWE Multilingual. The benchmark falls in the Coding category. BenchAlign v5.8 gives SWE Multilingual 5% of the Coding reference weight, so it moves the Coding leaderboard and the overall ranking. Reference weights are relative weights in the calibrated model, not fixed shares of a score.
About SWE Multilingual
Year
2026
Tasks
Multilingual software-engineering tasks
Format
Repository task completion
Difficulty
Professional software engineering
MiniMax reports SWE Multilingual as a coding benchmark focused on multilingual software-engineering tasks beyond single-language Python issue fixing.
Freshness and provenance
Version
SWE Multilingual 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does SWE Multilingual measure?
A multilingual software-engineering benchmark for real-world code issue resolution across multiple programming languages.
Which model scores highest on SWE Multilingual?
Claude Opus 5.5 by Anthropic currently leads with a score of 93.9% on SWE Multilingual.
How many models are evaluated on SWE Multilingual?
60 AI models have published results on SWE Multilingual in the BenchLM catalog.
Compare top models on SWE Multilingual
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 6,100+ subscribers.
One email each week. Unsubscribe anytime.