Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Instruction Following on CF-Bench

71Instruction Success Rate

Claude-Opus-4.7

12.7627.884358.12Feb 17, 2025May 2, 2025Jul 15, 2025Sep 27, 2025Dec 10, 2025Feb 22, 2026May 8, 2026
Updated 23d ago

Evaluation Results

MethodLinks
2026.05
71--
2026.05
69--
2026.05
68--
2026.05
68--
2026.05
68--
2026.05
66--
2026.05
65--
2026.05
60--
2026.05
59--
2026.05
58--
2026.05
55--
2026.05
52--
2026.05
51--
2026.05
50--
2026.05
49--
2026.05
47--
2025.02
447955
2026.05
43--
2025.02
427854
2025.02
417752
2026.05
41--
2025.02
407652
2025.02
387550
2025.02
387550
2025.02
377549
2025.02
377549
2026.05
37--
2025.02
367449
2026.05
36--
2025.02
357246
2026.05
35--
2026.05
34--
2026.05
34--
2025.02
327044
2025.02
306942
2025.02
296040
2025.02
296840
2025.02
296740
2025.02
286839
2025.02
286738
2025.02
286637
2025.02
276637
2025.02
266535
2025.02
266637
2025.02
256637
2026.05
25--
2026.05
25--
2025.02
246434
2026.01
246433
2026.05
24--
2026.05
24--
2025.02
236032
2026.01
236029
2026.01
236332
2026.01
236332
2026.05
23--
2026.01
226130
2026.01
226231
2026.01
226231
2026.05
22--
2026.01
216130
2026.01
216029
2025.02
205227
2025.02
205626
2026.01
195927
2026.01
185726
2026.01
175625
2026.01
155122