reasoning benchmark
SWE-Bench Verified Leaderboard
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Updated Aug 17, 2026
Models100
Model coverage111
MetricScore
EvidenceC
SWE-Bench Verified Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAN | Score95.0% | Percentile100.0% | Participants111 | EvidenceC | Evaluated |
| Rank02 | ModelAN | Score93.9% | Percentile99.1% | Participants111 | EvidenceC | Evaluated |
| Rank03 | ModelAN | Score88.6% | Percentile98.2% | Participants111 | EvidenceC | Evaluated |
| Rank04 | ModelAN | Score87.6% | Percentile97.3% | Participants111 | EvidenceC | Evaluated |
| Rank05 | ModelAN | Score85.2% | Percentile96.4% | Participants111 | EvidenceC | Evaluated |
| Rank06 | ModelAN | Score80.9% | Percentile95.5% | Participants111 | EvidenceC | Evaluated |
| Rank07 | ModelAN | Score80.8% | Percentile94.5% | Participants111 | EvidenceC | Evaluated |
| Rank08 | ModelDE | Score80.6% | Percentile93.6% | Participants111 | EvidenceC | Evaluated |
| Rank09 | ModelGO | Score80.6% | Percentile92.7% | Participants111 | EvidenceC | Evaluated |
| Rank10 | ModelMI | Score80.5% | Percentile91.8% | Participants111 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score80.4% | Percentile90.9% | Participants111 | EvidenceC | Evaluated |
| Rank12 | ModelTM | Score80.2% | Percentile90.0% | Participants111 | EvidenceC | Evaluated |
| Rank13 | ModelMA | Score80.2% | Percentile89.1% | Participants111 | EvidenceC | Evaluated |
| Rank14 | ModelMI | Score80.2% | Percentile88.2% | Participants111 | EvidenceC | Evaluated |
| Rank15 | ModelOP | Score80.0% | Percentile87.3% | Participants111 | EvidenceC | Evaluated |
| Rank16 | ModelAN | Score79.6% | Percentile86.4% | Participants111 | EvidenceC | Evaluated |
| Rank17 | ModelDE | Score79.0% | Percentile85.5% | Participants111 | EvidenceC | Evaluated |
| Rank18 | ModelXI | Score78.9% | Percentile84.5% | Participants111 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score78.8% | Percentile83.6% | Participants111 | EvidenceC | Evaluated |
| Rank20 | ModelDE | Score78.6% | Percentile82.7% | Participants111 | EvidenceC | Evaluated |
| Rank21 | ModelGO | Score78.0% | Percentile81.8% | Participants111 | EvidenceC | Evaluated |
| Rank22 | ModelTE | Score78.0% | Percentile80.9% | Participants111 | EvidenceC | Evaluated |
| Rank23 | ModelXI | Score78.0% | Percentile80.0% | Participants111 | EvidenceC | Evaluated |
| Rank24 | ModelZA | Score77.8% | Percentile79.1% | Participants111 | EvidenceC | Evaluated |
| Rank25 | ModelAC | Score77.7% | Percentile78.2% | Participants111 | EvidenceC | Evaluated |
| Rank26 | ModelMA | Score77.6% | Percentile77.3% | Participants111 | EvidenceC | Evaluated |
| Rank27 | ModelME | Score77.4% | Percentile76.4% | Participants111 | EvidenceC | Evaluated |
| Rank28 | ModelAC | Score77.2% | Percentile75.5% | Participants111 | EvidenceC | Evaluated |
| Rank29 | ModelMA | Score76.8% | Percentile74.5% | Participants111 | EvidenceC | Evaluated |
| Rank30 | ModelBY | Score76.5% | Percentile73.6% | Participants111 | EvidenceC | Evaluated |
| Rank31 | ModelAC | Score76.4% | Percentile72.7% | Participants111 | EvidenceC | Evaluated |
| Rank32 | ModelOP | Score76.3% | Percentile71.8% | Participants111 | EvidenceC | Evaluated |
| Rank33 | ModelOP | Score76.3% | Percentile70.9% | Participants111 | EvidenceC | Evaluated |
| Rank34 | ModelOP | Score76.3% | Percentile70.0% | Participants111 | EvidenceC | Evaluated |
| Rank35 | ModelGO | Score76.2% | Percentile69.1% | Participants111 | EvidenceC | Evaluated |
| Rank36 | ModelME | Score76.0% | Percentile68.2% | Participants111 | EvidenceC | Evaluated |
| Rank37 | ModelOP | Score74.9% | Percentile67.3% | Participants111 | EvidenceC | Evaluated |
| Rank38 | ModelXI | Score74.8% | Percentile66.4% | Participants111 | EvidenceC | Evaluated |
| Rank39 | ModelAN | Score74.5% | Percentile65.5% | Participants111 | EvidenceC | Evaluated |
| Rank40 | ModelOP | Score74.5% | Percentile64.5% | Participants111 | EvidenceC | Evaluated |
| Rank41 | ModelST | Score74.4% | Percentile63.6% | Participants111 | EvidenceC | Evaluated |
| Rank42 | ModelZA | Score73.8% | Percentile62.7% | Participants111 | EvidenceC | Evaluated |
| Rank43 | ModelOP | Score73.7% | Percentile61.8% | Participants111 | EvidenceC | Evaluated |
| Rank44 | ModelMI | Score73.5% | Percentile60.9% | Participants111 | EvidenceC | Evaluated |
| Rank45 | ModelBY | Score73.5% | Percentile60.0% | Participants111 | EvidenceC | Evaluated |
| Rank46 | ModelXI | Score73.4% | Percentile59.1% | Participants111 | EvidenceC | Evaluated |
| Rank47 | ModelAC | Score73.4% | Percentile58.2% | Participants111 | EvidenceC | Evaluated |
| Rank48 | ModelAN | Score73.3% | Percentile57.3% | Participants111 | EvidenceC | Evaluated |
| Rank49 | ModelDE | Score73.1% | Percentile56.4% | Participants111 | EvidenceC | Evaluated |
| Rank50 | ModelDE | Score73.1% | Percentile55.5% | Participants111 | EvidenceC | Evaluated |
| Rank51 | ModelDE | Score73.1% | Percentile54.5% | Participants111 | EvidenceC | Evaluated |
| Rank52 | ModelAN | Score72.7% | Percentile53.6% | Participants111 | EvidenceC | Evaluated |
| Rank53 | ModelMI | Score72.6% | Percentile52.7% | Participants111 | EvidenceC | Evaluated |
| Rank54 | ModelAN | Score72.5% | Percentile51.8% | Participants111 | EvidenceC | Evaluated |
| Rank55 | ModelAC | Score72.4% | Percentile50.9% | Participants111 | EvidenceC | Evaluated |
| Rank56 | ModelAC | Score72.0% | Percentile50.0% | Participants111 | EvidenceC | Evaluated |
| Rank57 | ModelMI | Score71.6% | Percentile49.1% | Participants111 | EvidenceC | Evaluated |
| Rank58 | ModelMA | Score71.3% | Percentile48.2% | Participants111 | EvidenceC | Evaluated |
| Rank59 | ModelPO | Score70.9% | Percentile47.3% | Participants111 | EvidenceC | Evaluated |
| Rank60 | ModelXA | Score70.8% | Percentile46.4% | Participants111 | EvidenceC | Evaluated |
| Rank61 | ModelNV | Score70.7% | Percentile45.5% | Participants111 | EvidenceC | Evaluated |
| Rank62 | ModelUP | Score70.6% | Percentile44.5% | Participants111 | EvidenceC | Evaluated |
| Rank63 | ModelAN | Score70.3% | Percentile43.6% | Participants111 | EvidenceC | Evaluated |
| Rank64 | ModelME | Score70.0% | Percentile42.7% | Participants111 | EvidenceC | Evaluated |
| Rank65 | ModelAM Nova 2 ProAmazon | Score70.0% | Percentile41.8% | Participants111 | EvidenceC | Evaluated |
| Rank66 | ModelAC | Score69.6% | Percentile40.9% | Participants111 | EvidenceC | Evaluated |
| Rank67 | ModelAC | Score69.6% | Percentile40.0% | Participants111 | EvidenceC | Evaluated |
| Rank68 | ModelMI | Score69.4% | Percentile39.1% | Participants111 | EvidenceC | Evaluated |
| Rank69 | ModelAC | Score69.2% | Percentile38.2% | Participants111 | EvidenceC | Evaluated |
| Rank70 | ModelOP | Score69.1% | Percentile37.3% | Participants111 | EvidenceC | Evaluated |
| Rank71 | ModelOP | Score68.1% | Percentile36.4% | Participants111 | EvidenceC | Evaluated |
| Rank72 | ModelZA | Score68.0% | Percentile35.5% | Participants111 | EvidenceC | Evaluated |
| Rank73 | ModelDE | Score67.8% | Percentile34.5% | Participants111 | EvidenceC | Evaluated |
| Rank74 | ModelCO | Score67.6% | Percentile33.6% | Participants111 | EvidenceC | Evaluated |
| Rank75 | ModelGO | Score67.2% | Percentile32.7% | Participants111 | EvidenceC | Evaluated |
| Rank76 | ModelMI | Score67.0% | Percentile31.8% | Participants111 | EvidenceC | Evaluated |
| Rank77 | ModelDE | Score66.0% | Percentile30.9% | Participants111 | EvidenceC | Evaluated |
| Rank78 | ModelMA | Score65.8% | Percentile30.0% | Participants111 | EvidenceC | Evaluated |
| Rank79 | ModelAMNova 2 LiteAmazon | Score64.5% | Percentile29.1% | Participants111 | EvidenceC | Evaluated |
| Rank80 | ModelZA | Score64.2% | Percentile28.2% | Participants111 | EvidenceC | Evaluated |
| Rank81 | ModelGO | Score63.2% | Percentile27.3% | Participants111 | EvidenceC | Evaluated |
| Rank82 | ModelMA | Score61.6% | Percentile26.4% | Participants111 | EvidenceC | Evaluated |
| Rank83 | ModelGO | Score60.4% | Percentile25.4% | Participants111 | EvidenceC | Evaluated |
| Rank84 | ModelME | Score60.4% | Percentile24.6% | Participants111 | EvidenceC | Evaluated |
| Rank85 | ModelME | Score59.4% | Percentile23.6% | Participants111 | EvidenceC | Evaluated |
| Rank86 | ModelZA | Score59.2% | Percentile22.7% | Participants111 | EvidenceC | Evaluated |
| Rank87 | ModelZA | Score57.6% | Percentile21.8% | Participants111 | EvidenceC | Evaluated |
| Rank88 | ModelMI | Score56.0% | Percentile20.9% | Participants111 | EvidenceC | Evaluated |
| Rank89 | ModelMI | Score55.6% | Percentile20.0% | Participants111 | EvidenceC | Evaluated |
| Rank90 | ModelOP | Score54.6% | Percentile19.1% | Participants111 | EvidenceC | Evaluated |
| Rank91 | ModelME | Score54.4% | Percentile18.2% | Participants111 | EvidenceC | Evaluated |
| Rank92 | ModelNV | Score53.7% | Percentile17.3% | Participants111 | EvidenceC | Evaluated |
| Rank93 | ModelMA | Score53.6% | Percentile16.4% | Participants111 | EvidenceC | Evaluated |
| Rank94 | ModelNV | Score51.6% | Percentile15.4% | Participants111 | EvidenceC | Evaluated |
| Rank95 | ModelOP | Score49.3% | Percentile14.6% | Participants111 | EvidenceC | Evaluated |
| Rank96 | ModelAN | Score49.0% | Percentile13.6% | Participants111 | EvidenceC | Evaluated |
| Rank97 | ModelSA | Score45.0% | Percentile12.7% | Participants111 | EvidenceC | Evaluated |
| Rank98 | ModelDE | Score44.6% | Percentile11.8% | Participants111 | EvidenceC | Evaluated |
| Rank99 | ModelDE | Score42.0% | Percentile10.9% | Participants111 | EvidenceC | Evaluated |
| Rank100 | ModelOP | Score41.3% | Percentile10.0% | Participants111 | EvidenceC | Evaluated |
SWE-Bench Verified Highlights
The leading models and scores on this benchmark.
SWE-Bench Verified Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for SWE-Bench Verified
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis swe-bench verified AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01ANClaude Fable 5Anthropic
- Score
- 95.0%
- Price
- $10 input / $50 output per 1M tokens
- Speed
- Up to 9.1 tok/s via Anthropic
Strengths
- Ranks #1 of 111 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures SWE-Bench Verified, not total model capability
- 02ANClaude Mythos PreviewAnthropic
- Score
- 93.9%
Strengths
- Ranks #2 of 111 compared models
- 99th percentile on this benchmark
- C evidence result
Considerations
- This result measures SWE-Bench Verified, not total model capability
- 03ANClaude Opus 4.8Anthropic
- Score
- 88.6%
- Price
- $5.0 input / $25 output per 1M tokens
- Speed
- Up to 46 tok/s via Anthropic
Strengths
- Ranks #3 of 111 compared models
- 98th percentile on this benchmark
- C evidence result
Considerations
- This result measures SWE-Bench Verified, not total model capability
- 04ANClaude Opus 4.7Anthropic
- Score
- 87.6%
- Price
- $5.0 input / $25 output per 1M tokens
- Speed
- Up to 42 tok/s via Anthropic
Strengths
- Ranks #4 of 111 compared models
- 97th percentile on this benchmark
- C evidence result
Considerations
- This result measures SWE-Bench Verified, not total model capability
- 05ANClaude Sonnet 5Anthropic
- Score
- 85.2%
- Price
- $2.0 input / $10 output per 1M tokens
- Speed
- Up to 11 tok/s via Anthropic
Strengths
- Ranks #5 of 111 compared models
- 96th percentile on this benchmark
- C evidence result
Considerations
- This result measures SWE-Bench Verified, not total model capability
Selection summary
Best AI Models for SWE-Bench Verified
Claude Fable 5 currently leads SWE-Bench Verified with 95.0%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is SWE-Bench Verified?
What SWE-Bench Verified measures and how its scores work.
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of C.
- Family
- SWE-Bench Verified
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- swe-bench-verified|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about SWE-Bench Verified.
Which model scores highest on SWE-Bench Verified?
Claude Fable 5 is currently ranked first with 95.0%.
What does SWE-Bench Verified measure?
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
100 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.