context.vn

swe Multilingual

30 models evaluated

#ModelProviderTypeScore
1Claude Opus 5AnthropicClosedclaude-opus-5
2Claude Opus 4.8AnthropicClosedclaude-opus-4-8
3Composer 2.5CursorClosedcomposer-2-5
4Ornith-1.0-397BDeepReinforce AI · Open weight78.9%
5Laguna S 2.1Poolside · Open weight78.5%
6Qwen3.7 MaxAlibaba · Closed78.3%
7Claude Sonnet 5Anthropic · Closed78.3%
8Grok 4.5xAI · Closed78%
9SWE-1.7Cognition · Closed77.8%
10Claude Opus 4.5Anthropic · Closed77.5%
11Kimi K2.6Moonshot AI · Open weight76.7%
12MiniMax M2.7MiniMax · Open weight76.5%
13DeepSeek V4 Pro (Max)DeepSeek · Open weight76.2%
14Qwen3.7 PlusAlibaba · Closed75.8%
15DeepSeek V4 Pro (High)DeepSeek · Open weight74.1%
16Qwen3.6 PlusAlibaba · Closed73.8%
17Composer 2Cursor · Closed73.7%
18GLM-5Z.AI · Open weight73.3%
19DeepSeek V4 Flash (Max)DeepSeek · Open weight73.3%
20Kimi K2.5Moonshot AI · Open weight73%
21Qwen3.6-27BAlibaba · Open weight71.3%
22DeepSeek V4 Flash (High)DeepSeek · Open weight70.2%
23DeepSeek V4 ProDeepSeek · Open weight69.8%
24DeepSeek V4 FlashDeepSeek · Open weight69.7%
25Ornith-1.0-35BDeepReinforce AI · Open weight69.3%
26Nemotron 3 UltraNVIDIA · Open weight67.7%
27Qwen3.6-35B-A3BAlibaba · Open weight67.2%
28Laguna M.1Poolside · Closed63.1%
29Laguna XS.2Poolside · Open weight57.7%
30Ornith-1.0-9BDeepReinforce AI · Open weight52%