SWE-bench team

SWE-bench Verified · swebench.com

Resolved-issue rate on real GitHub issues, as published by the SWE-bench team.

In index
Coding
Source ↗

Models46

Top result74.4%Claude 4.5 Opus

Top-3 spread3.8 pts

Updated26 Feb 2026

Result and cost

Results

46 results

#ModelResultCost

01

Claude 4.5 Opusmedium effort · mini-SWE-agent74.4%$0.721

02

Gemini 3 Promini-SWE-agent74.2%$0.46

03

Claude 4.5 Sonnetmini-SWE-agent70.6%$0.558

04

Claude 4 SonnetOpenHands68.5%—

05

Claude 4 Opusmini-SWE-agent67.6%$1.13

06MAKimi K2OpenHands65.4%—

07

GPT-5medium effort · mini-SWE-agent65.0%$0.28

08MAKimi K2 Thinkingmini-SWE-agent63.4%$0.438

09

Claude 3.7 Sonnet w/ Review HeavySWE-agent62.4%—

10

MiniMax M2mini-SWE-agent61.0%$0.428

11UN4x ScaledOpenHands60.8%—

12

DeepSeek V3.2 Reasonermini-SWE-agent60.0%$0.028

13

GPT-5 minimedium effort · mini-SWE-agent59.8%$0.035

14

o3mini-SWE-agent58.4%$0.334

15

Devstral Small (2512)mini-SWE-agent56.4%—

16

GLM-4.6mini-SWE-agent55.4%$0.097

16

Qwen3-Coder 480B/A35B Instructmini-SWE-agent55.4%$0.248

18

GLM-4.5mini-SWE-agent54.2%$0.297

19

Devstral 2 2512mini-SWE-agent53.8%—

20

Gemini 2.5 Promini-SWE-agent53.6%$0.288

21UNCodeAct v2.1 (claude-3-5-sonnet-20241022)OpenHands53.0%—

22

Claude 3.7 Sonnetmini-SWE-agent52.8%$0.355

—

DevStral Small 2505OpenHands46.8%—

23

o4-mini (high)mini-SWE-agent45.0%$0.21

24MAKimi K2 Instructmini-SWE-agent43.8%$0.532

25

o3-miniAgentless Lite42.4%—

26UNSWE Agent LM 32BSWE-agent40.2%—

27

GPT-4.1mini-SWE-agent39.6%$0.146

28

DevStral Small 2507SWE-agent38.0%—

29

GPT-5 nanomedium effort · mini-SWE-agent34.8%$0.038

30

Claude 3.5 SonnetSWE-agent33.6%—

31

Qwen2.5 (7B + 72B)SWE-Fixer32.8%—

32

Gemini 2.5 Flashmini-SWE-agent28.7%$0.134

33

GPT-OSS 120Bmini-SWE-agent26.0%$0.057

34

GPT-4.1 minimini-SWE-agent23.9%$0.435

35

Qwen2.5 Coder 32B InstructSkywork-SWE-32B23.5%—

36

GPT-4oSWE-agent23.2%—

37

Llama 4 Maverick Instructmini-SWE-agent21.0%$0.314

38

Gemini 2.0 Flashmini-SWE-agent13.5%$0

39

GPT-4 (1106)SWE-agent12.6%—

40

Claude 3 OpusSWE-agent11.4%—

41

Llama 4 Scout Instructmini-SWE-agent9.1%$0.116

42

Claude 2RAG baseline4.4%—

43UNSWE-Llama 7BRAG baseline1.4%—

44UNSWE-Llama 13BRAG baseline1.2%—

45

GPT-3.5RAG baseline0.4%—