Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Agentic Coding on DevEval
Loading...
94.8
Solve Rate
GPT-5.4
70.464
76.782
83.1
89.418
Jun 11, 2026
Solve Rate
Average Cost per Instruction
Updated 1mo ago
Evaluation Results
Method
Method
Links
Solve Rate
Average Cost per Instruction
GPT-5.4
Harness=Codex, Thinkin...
2026.06
94.8
0.2
Claude Opus 4.7
Harness=Claude Code, T...
2026.06
91.9
0.26
Gemini 3.1 Pro
Harness=OpenCode, Thin...
2026.06
88.8
0.52
Qwen3.5-397B-A17B
Harness=MSA, Thinking=...
2026.06
71.4
-
Feedback
Search any
task
Search any
task