Edison Advances
Benchmarks / LABBench2

FigQA2 (pdf)

figqa2-pdf

10 runs · 5 models · evaluated by Score.

# Model ↕ Variant ↕ Mode ↕ Average score ↓ Average output tokens ↕ Avg. dur ↕ Date ↕
1 gpt-5-2 tools,high file 0.644 6.1k 11.2m 2026-02-03
2 gpt-5-2-pro tools,high file 0.436 2.9k 7.0m 2026-02-03
3 gemini-3-pro-preview — file 0.416 202 1.7m 2026-02-03
4 claude-opus-4-6 — file 0.396 219 36.6s 2026-03-20
5 gpt-5-2-pro — file 0.396 1.1k 3.1m 2026-02-03
6 claude-opus-4-5 tools,high file 0.386 2.7k 55.3s 2026-03-22
7 claude-opus-4-6 tools,high file 0.366 4.9k 2.2m 2026-03-22
8 gpt-5-2 — file 0.366 101 1.5m 2026-02-03
9 claude-opus-4-5 — file 0.356 302 22.4s 2026-03-20
10 gemini-3-pro-preview tools,high file 0.337 1.5k 3.7m 2026-02-03

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