Edison Labs
Benchmarks / LabBench2

Cloning

cloning

32 runs · 5 models · evaluated by HybridEvaluator.

Mode
# Model Variant Mode Score Avg. dur Tokens Date
1 gemini-3-pro-preview tools,high file 0.429 2.6m 283.2k 2026-01-27
2 claude-opus-4-6 tools,high inject 0.429 21.5m 10.7M 2026-03-22
3 claude-opus-4-6 tools,high file 0.357 14.3m 12.7M 2026-03-22
4 gemini-3-pro-preview file 0.357 1.7m 231.7k 2026-01-27
5 gemini-3-pro-preview inject 0.357 2.1m 229.5k 2026-01-22
6 gemini-3-pro-preview tools,high inject 0.357 3.3m 334.4k 2026-01-23
7 claude-opus-4-5 tools,high inject 0.286 10.4m 6.4M 2026-03-22
8 gpt-5-2-pro inject 0.286 9.8m 392.2k 2026-01-22
9 gpt-5-2 tools,high inject 0.286 14.0m 871.4k 2026-01-23
10 claude-opus-4-5 tools,high retrieve 0.286 5.5m 10.7M 2026-03-22
11 claude-opus-4-6 tools,high retrieve 0.286 17.9m 21.9M 2026-03-22
12 gpt-5-2 file 0.214 20.2s 223.1k 2026-01-28
13 claude-opus-4-6 inject 0.214 1.8m 329.7k 2026-03-20
14 gpt-5-2-pro tools,high inject 0.214 14.8m 138.3k 2026-01-25
15 gpt-5-2 inject 0.214 20.0s 224.1k 2026-01-22
16 claude-opus-4-5 retrieve 0.214 24.8s 25.4k 2026-03-20
17 claude-opus-4-6 retrieve 0.214 28.2s 26.2k 2026-03-20
18 gemini-3-pro-preview retrieve 0.214 52.9s 18.0k 2026-01-26
19 gpt-5-2-pro retrieve 0.214 5.1m 110.4k 2026-01-26
20 gpt-5-2 tools,high retrieve 0.214 9.7m 876.8k 2026-01-26
21 claude-opus-4-5 file 0.143 27.1s 1.7M 2026-03-20
22 claude-opus-4-5 tools,high file 0.143 7.5m 13.1M 2026-03-22
23 gpt-5-2 tools,high file 0.143 12.7m 507.9k 2026-01-28
24 claude-opus-4-5 inject 0.143 26.0s 265.3k 2026-03-20
25 gpt-5-2-pro tools,high retrieve 0.143 13.7m 566.9k 2026-01-26
26 claude-opus-4-6 file 0.071 40.0s 2.5M 2026-03-20
27 gemini-3-pro-preview tools,high retrieve 0.071 1.8m 26.5k 2026-01-26
28 gpt-5-2-pro file 0.0s 0 2026-01-28
29 gpt-5-2-pro tools,high file 0.0s 0 2026-01-23
30 gpt-5-2-pro tools,high_retry inject 8.7m 23.1k 2026-01-25
31 gpt-5-2-pro tools,high_retry_retry inject 9.5m 42.9k 2026-01-25
32 gpt-5-2 retrieve 52.6s 17.1k 2026-01-27

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