Edison Advances
Benchmarks / LABBench2

SourceQuality

sourcequality

10 runs · 5 models · evaluated by Score.

# Model ↕ Variant ↕ Mode ↕ Average score ↓ Average output tokens ↕ Avg. dur ↕ Date ↕
1 gemini-3-pro-preview code,high file 0.900 256 45.1s 2026-03-13
2 gemini-3-pro-preview — file 0.887 238 35.6s 2026-03-13
3 gpt-5-2-pro — file 0.807 480 1.6m 2026-03-13
4 gpt-5-2 — file 0.793 163 29.9s 2026-03-13
5 gpt-5-2-pro code,high file 0.760 923 2.6m 2026-03-13
6 gpt-5-2 code,high file 0.720 3.5k 3.1m 2026-03-13
7 claude-opus-4-6 code,high file 0.673 1.9k 1.3m 2026-03-23
8 claude-opus-4-5 — file 0.660 326 16.5s 2026-03-20
9 claude-opus-4-5 code,high file 0.620 1.1k 30.2s 2026-03-23
10 claude-opus-4-6 — file 0.607 261 19.0s 2026-03-20

Click column headers to sort. Click mode chips to filter.