[verbatim] FF1 all                      Claude vs Gemini: 534 blocks | means 1.547 vs 1.918 | r_rb -0.396 [-0.555, -0.228] | p_holm 4.57e-07
[verbatim] FF1 all                      Claude vs GPT: 534 blocks | means 1.547 vs 2.453 | r_rb -0.646 [-0.763, -0.512] | p_holm < 2.2e-16
[verbatim] FF1 all                      Gemini vs GPT: 534 blocks | means 1.918 vs 2.453 | r_rb -0.390 [-0.533, -0.232] | p_holm 4.57e-07
[verbatim] FF1 V1_blocks_without_copy   Claude vs Gemini: 494 blocks | means 1.621 vs 1.988 | r_rb -0.381 [-0.545, -0.208] | p_holm 1.23e-06
[verbatim] FF1 V1_blocks_without_copy   Claude vs GPT: 494 blocks | means 1.621 vs 2.573 | r_rb -0.643 [-0.761, -0.508] | p_holm 9.02e-16
[verbatim] FF1 V1_blocks_without_copy   Gemini vs GPT: 494 blocks | means 1.988 vs 2.573 | r_rb -0.410 [-0.552, -0.253] | p_holm 2.88e-07
[verbatim] FF1 V2_tasks_without_copy    Claude vs Gemini: 489 blocks | means 1.618 vs 1.975 | r_rb -0.372 [-0.542, -0.197] | p_holm 2.44e-06
[verbatim] FF1 V2_tasks_without_copy    Claude vs GPT: 489 blocks | means 1.618 vs 2.579 | r_rb -0.643 [-0.763, -0.508] | p_holm 9.02e-16
[verbatim] FF1 V2_tasks_without_copy    Gemini vs GPT: 489 blocks | means 1.975 vs 2.579 | r_rb -0.428 [-0.570, -0.269] | p_holm 1.05e-07
[verbatim] reference (Sonar) without copies Claude  n = 663 | mean gen 1.84 vs ref 2.42, diff -0.58 [-0.86, -0.32] | r_rb -0.386 [-0.532, -0.232] | p_holm 2.37e-09
[verbatim] reference (Sonar) without copies Gemini  n = 600 | mean gen 2.06 vs ref 2.29, diff -0.22 [-0.50, +0.05] | r_rb -0.151 [-0.305, +0.012] | p_holm 3.31e-02
[verbatim] reference (Sonar) without copies GPT     n = 577 | mean gen 2.71 vs ref 2.38, diff +0.33 [-0.01, +0.68] | r_rb +0.121 [-0.037, +0.280] | p_holm 5.10e-02
