SWE Atlas - Codebase QnA

Kimi K3
Kimi K3Max
52.7%
Opus 5
Opus 5Max
50.3%
Opus 4.7
Opus 4.7Max
47.0%
GLM-5.2
GLM-5.2Max
46.0%
GPT-5.6 Luna
GPT-5.6 LunaMax
46.0%
GPT-5.5
GPT-5.5xHigh
45.2%
GPT-5.6 Terra
GPT-5.6 TerraMax
45.2%
DeepSeek-V4-Flash
DeepSeek-V4-FlashMax
44.9%
Opus 4.8
Opus 4.8Max
44.4%
Fable 5.1
Fable 5.1High
42.7%
Grok 4.5
Grok 4.5High
41.9%
Sonnet 5
Sonnet 5Max
39.5%
GPT-5.4
GPT-5.4xHigh
37.1%
Sonnet 4.6
Sonnet 4.6High
35.5%
Kimi K2.7 Code
Kimi K2.7 CodeHigh
33.1%
Fable 5
Fable 5Max
33.1%
Gemini 3.6 Flash
Gemini 3.6 FlashHigh
32.0%
DeepSeek-V4-Pro
DeepSeek-V4-ProMax
25.0%
Inkling
InklingHigh
21.2%
Gemini 3.5 Flash
Gemini 3.5 FlashHigh
19.6%
Gemini 3.1 Pro
Gemini 3.1 ProHigh
18.0%
DeepSeek-V3.2
DeepSeek-V3.2
5.9%
0%
10%
20%
30%
40%
50%
60%
70%

APEX NEWSLETTER

The latest on frontier AI performance, straight to your inbox.

New benchmarks, leaderboard shifts, and research from the APEX team.

By subscribing you agree to receive updates from Mercor.
Unsubscribe anytime.