LLM Reference
activeCoding

SWE-rebench

Metric: Resolved Rate (higher is better)Introduced: 2025

Evaluates LLM coding agents on real-world GitHub issues sourced after each model's training cutoff, preventing benchmark contamination. Uses standardized ReAct scaffolding with 128K token context; each model is run five times per problem and the best Pass@1 resolved rate is reported.

Models ranked

14

tracked on this benchmark

Score band

79.8 – 41.6

best → lowest tracked

Snapshot trend

+21.96

May 28 → Jun 7 · 1 models

Leaderboard

Tracked models ranked by Resolved Rate (higher is better).

Compare candidates
#Model variant and provenanceScore
1
Composer 2.5
Version: Reported as SWE-bench Multilingual (300 tasks, 9 languages) (resolved%)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
79.8
2
Claude Opus 4.6
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
65.3
3
GLM-5
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
62.8
4
GLM-5.1
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
62.7
5
DeepSeek V3.2
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
60.9
6
Claude Sonnet 4.6
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
60.7
7
Qwen3.5-27B
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
58.9
8
GLM 4.7
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
58.7
9
Kimi K2.5
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
58.5
10
GPT-5.3-Codex
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
58.2
11
Composer 2
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
58.0
12
Qwen3.5-35B-A3B
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
53.7
13
MiniMax M2.7
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
51.9
14
Gemma 4 31B
Version: pass@1 (best of 5 runs)Harness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource
41.6

How to read this benchmark

This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.

Trust this score when

  • There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
  • The model list covers the same version family you can actually deploy today.
  • Top candidates overlap with your required routing and feature requirements.

Be cautious when

  • There is only one benchmark snapshot or the dataset appears stale.
  • The benchmark metric direction is opposite of your decision objective.
  • The score difference between options is narrow and likely within implementation variance.

Related benchmarks

Last reviewed: May 28, 2026