Skip to content

Agents · long tasks · reference

SWE-Bench Pro

Longer, harder software tasks across several languages, from Scale AI. Closer to a day of real engineering work than a single bug fix.

Models
24
By makers
24
Latest
Oct 13
#ModelBarScoreSource
1Claude Opus 5.5Anthropic · $8.00/1M · Coming soon89.9%vSep 22, 2026MakerSep 22, 2026
2Claude Sonnet 5.5Anthropic · $4.00/1M · Coming soon81.3%vSep 28, 2026MakerSep 28, 2026
3Claude Fable 5Anthropic · $20.00/1M · Coming soon80%vJun 9, 2026MakerJun 9, 2026
4Claude Opus 5Anthropic · $10.00/1M · Coming soon79.2%vJul 24, 2026MakerJul 24, 2026
5Claude Opus 4.8Anthropic · $10.00/1M · Coming soon69.2%vMay 28, 2026MakerMay 28, 2026
6Ling 3.1 FlashinclusionAI · $0.45/1M · Coming soon68.7%vOct 13, 2026MakerOct 13, 2026
7Qwen3.8 MaxQwen · $2.48/1M · Coming soon67.7%vAug 8, 2026MakerAug 8, 2026
8Claude Opus 4.7Anthropic · $10.00/1M · Coming soon64.3%vApr 16, 2026MakerApr 16, 2026
9Claude Sonnet 5Anthropic · $6.00/1M · Coming soon63.2%vJun 30, 2026MakerJun 30, 2026
10Qwen3.8 FlashQwen · $0.18/1M · Coming soon62.5%vAug 24, 2026MakerAug 24, 2026
11GLM 5.2Z.ai · $1.16/1M · Coming soon62.1%vJun 16, 2026MakerJun 16, 2026
12Qwen3.8 27BQwen · $0.78/1M · Coming soon61.7%vAug 5, 2026MakerAug 5, 2026
13Qwen3.7 MaxQwen · $3.75/1M · Coming soon60.6%vAug 8, 2026MakerAug 8, 2026
14MiniMax M3MiniMax · $0.48/1M · Coming soon59%vJun 2, 2026MakerJun 2, 2026
15Kimi K2.6Moonshot AI · $1.44/1M · Coming soon58.6%vApr 14, 2026MakerApr 14, 2026
16Inkling Small$0.64/1M · Coming soon55.9%vundatedMakerundated
17Gemini 3.5 FlashGoogle · $3.38/1M · Coming soon55.1%vMay 19, 2026MakerMay 19, 2026
18Gemini 3.1 ProGoogle · $4.50/1M · Coming soon54.2%vFeb 19, 2026MakerFeb 19, 2026
19Qwen3.6 27BQwen · $1.04/1M · Coming soon53.5%vApr 21, 2026MakerApr 21, 2026
20DeepSeek V4 FlashDeepSeek · $0.11/1M · Coming soon52.6%vApr 22, 2026MakerApr 22, 2026
21Muse Glimmer 30BMeta · $0.53/1M · Coming soon51.2%vundatedMakerundated
22Qwen3.6 35B A3BQwen · $0.31/1M · Coming soon49.5%vApr 15, 2026MakerApr 15, 2026
23Seed 2.0 ProByteDance · $1.13/1M · Coming soon46.9%vFeb 15, 2026MakerFeb 15, 2026
24Granite 4.2 8BIBM · $0.11/1M · Coming soon19.1%vAug 25, 2026MakerAug 25, 2026

% of tasks resolved. v marks a score the model’s maker reported. Prices are FullsendAI’s per million tokens, blended 3 input : 1 output.

About the data

From Scale AI SWE-Bench Pro. No explicit licence on the leaderboard data; Scale's general site terms apply. The benchmark code is MIT.

All benchmarksModel rankingsCompare models