Skip to main content

Benchmark profile

SWE-Rebench

A continuously updated software engineering benchmark by Nebius using fresh GitHub issues to avoid contamination. Models are evaluated 5 times per problem under a fixed ReAct scaffolding; the Resolved Rate (best pass@1) is reported.

Data verified

Claude Opus 4.6 leads the SWE-Rebench leaderboard on BenchLM's July 2026 update with 65.3%, ahead of GLM-5 (62.8%) and GLM-5.1 (62.7%), across 13 tracked models.

Top models on SWE-Rebench — July 23, 2026

As of July 23, 2026, Claude Opus 4.6 leads the SWE-Rebench leaderboard with 65.3% , followed by GLM-5 (62.8%) and GLM-5.1 (62.7%).

13 modelsCoding20% of category scoreCurrentUpdated July 23, 2026

Leaderboard (13 models)

Score
1
Claude Opus 4.6Anthropic · Closed
65.3%
2
GLM-5Z.AI · Open weight
62.8%
3
GLM-5.1Z.AI · Open weight
62.7%
4
DeepSeek V3.2DeepSeek · Open weight
60.9%
5
Claude Sonnet 4.6Anthropic · Closed
60.7%
6
Qwen3.5-27BAlibaba · Open weight
58.9%
7
GLM-4.7Z.AI · Open weight
58.7%
8
Kimi K2.5Moonshot AI · Open weight
58.5%
9
GPT-5.3 CodexOpenAI · Closed
58.2%
10
Composer 2Cursor · Closed
58%
11
Qwen3.5-35B-A3BAlibaba · Open weight
53.7%
12
MiniMax M2.7MiniMax · Open weight
51.9%
13
Gemma 4 31BGoogle · Open weight
41.6%

According to BenchLM.ai, Claude Opus 4.6 leads the SWE-Rebench benchmark with a score of 65.3%, followed by GLM-5 (62.8%) and GLM-5.1 (62.7%). The top models are clustered within 2.6 points, suggesting this benchmark is nearing saturation for frontier models.

13 models have been evaluated on SWE-Rebench. The benchmark falls in the Coding category. This category carries a 20% weight in BenchLM.ai's overall scoring system. Within that category, SWE-Rebench contributes 20% of the category score, so strong performance here directly affects a model's overall ranking.

About SWE-Rebench

Year

2026

Tasks

Fresh GitHub issues (rolling window)

Format

Code patch generation

Difficulty

Professional software engineering

SWE-Rebench uses a rolling window of fresh problems from real GitHub repositories, sourced after each model's release date to prevent contamination. Each model runs 5 times with a standardized 128K-context ReAct scaffold. Unlike SWE-bench Verified (2023 problems), scores reflect consistent, up-to-date difficulty.

BenchLM freshness & provenance

Version

Rolling 2026 window

Refresh cadence

Rolling

Staleness state

Current

Question availability

Rolling public issues

Current

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.

FAQ

What does SWE-Rebench measure?

A continuously updated software engineering benchmark by Nebius using fresh GitHub issues to avoid contamination. Models are evaluated 5 times per problem under a fixed ReAct scaffolding; the Resolved Rate (best pass@1) is reported.

Which model scores highest on SWE-Rebench?

Claude Opus 4.6 by Anthropic currently leads with a score of 65.3% on SWE-Rebench.

How many models are evaluated on SWE-Rebench?

13 AI models have been evaluated on SWE-Rebench on BenchLM.

Last updated: July 23, 2026 · BenchLM version Rolling 2026 window

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.