Scorecards Reports Results Disagreements Flags Reliability Runs 276 scored runs across 28 curricula
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge ss-experiment gemini-2.5-pro api 2026-06-01 85%n=34 0.56 100% 5 ss-experiment gemini-2.5-flash api 2026-06-01 76%n=34 0.54 67% 8 ss-experiment claude-sonnet-4-6 cli 2026-06-01 82%n=34 0.67 100% 6 ss-experiment claude-haiku-4-5 cli 2026-06-01 62%n=34 0.32 100% 13 timing-vision claude-opus-4-7 api 2026-05-31 100%n=6 1.00 100% 0 Signal Studio claude-opus-4-8 cli 2026-05-31 82%n=34 0.49 100% 6 vision-img claude-opus-4-7 api 2026-05-31 82%n=34 0.72 100% 6 vision-base claude-opus-4-7 api 2026-05-31 82%n=34 0.72 100% 6 Signal Studio claude-opus-4-7 cli 2026-05-28 82%n=34 0.67 100% 6
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 100%n=29 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=29 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=29 1.00 100% 0 Signal Studio claude-sonnet-4-6 cli 2026-05-28 100%n=29 1.00 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 93%n=29 0.94 100% 2 Signal Studio claude-opus-4-7 cli 2026-05-26 100%n=29 1.00 100% 0 AutoEdReport — 2026-05-26 100%n=29 1.00 100% 0 AutoEdReport — 2026-05-26 100%n=29 1.00 100% 0 AutoEdReport — 2026-05-26 100%n=29 1.00 100% 0 AutoEdReport — 2026-05-26 97%n=29 0.97 100% 1 AutoEdReport — 2026-05-26 31%n=29 0.27 0% 20 AutoEdReport — 2026-05-26 48%n=29 0.28 0% 15 AutoEdReport — 2026-05-26 97%n=29 0.97 100% 1 Signal Studio claude-sonnet-4-6 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-sonnet-4-6 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-opus-4-7 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-opus-4-7 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-sonnet-4-6 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-opus-4-7 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-sonnet-4-6 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-haiku-4-5 2026-05-25 100%n=29 1.00 100% 0 Signal Studio claude-sonnet-4-6 2026-05-24 90%n=29 0.80 100% 3 Signal Studio claude-sonnet-4-6 2026-05-24 100%n=29 1.00 100% 0 Signal Studio claude-sonnet-4-6 2026-05-24 97%n=29 0.86 100% 1 Signal Studio claude-sonnet-4-6 2026-05-24 100%n=29 1.00 100% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 100%n=17 1.00 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 71%n=17 0.00 50% 5 Signal Studio claude-sonnet-4-6 cli 2026-05-28 100%n=17 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=17 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 100%n=17 1.00 100% 0 AutoEdReport — 2026-05-27 94%n=17 0.00 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 100%n=19 1.00 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 84%n=19 0.49 67% 3 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=19 1.00 100% 0 Signal Studio claude-sonnet-4-6 cli 2026-05-28 100%n=19 1.00 100% 0 AutoEdReport — 2026-05-27 84%n=19 0.73 100% 3 Signal Studio claude-opus-4-7 cli 2026-05-27 95%n=19 0.89 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 100%n=17 1.00 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 59%n=17 0.00 50% 7 Signal Studio claude-sonnet-4-6 cli 2026-05-28 88%n=17 0.00 100% 2 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=17 1.00 100% 0 AutoEdReport — 2026-05-27 88%n=17 0.00 100% 2 Signal Studio claude-opus-4-7 cli 2026-05-27 100%n=17 1.00 100% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 100%n=16 1.00 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 56%n=16 0.00 50% 7 Signal Studio claude-sonnet-4-6 cli 2026-05-28 100%n=16 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=16 1.00 100% 0 AutoEdReport — 2026-05-27 100%n=16 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 100%n=16 1.00 100% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-31 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-31 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-31 95%n=19 0.89 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-31 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-31 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-31 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-29 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-29 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-29 100%n=19 1.00 100% 0 ss-experiment claude-opus-4-7 cli 2026-05-29 95%n=19 0.89 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-29 100%n=19 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=19 1.00 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 84%n=19 0.58 67% 3 Signal Studio claude-sonnet-4-6 cli 2026-05-28 100%n=19 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 100%n=19 1.00 100% 0 AutoEdReport — 2026-05-27 84%n=19 0.73 100% 3 Signal Studio claude-opus-4-7 cli 2026-05-27 95%n=19 0.89 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-31 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-31 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-31 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-31 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-29 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-29 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-29 97%n=30 0.90 100% 1 ss-experiment claude-opus-4-7 cli 2026-05-29 97%n=30 0.90 100% 1 Signal Studio claude-opus-4-7 cli 2026-05-28 97%n=30 0.90 100% 1 Signal Studio claude-haiku-4-5 cli 2026-05-28 79%n=29 0.73 100% 6 Signal Studio claude-sonnet-4-6 cli 2026-05-28 97%n=30 0.90 100% 1 Signal Studio claude-opus-4-7 cli 2026-05-28 97%n=30 0.90 100% 1 Signal Studio claude-opus-4-7 cli 2026-05-26 97%n=30 0.90 100% 1 AutoEdReport — 2026-05-26 97%n=30 0.97 100% 1 Signal Studio claude-opus-4-7 2026-05-25 97%n=30 0.90 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 95%n=22 0.89 100% 1 Signal Studio claude-haiku-4-5 cli 2026-05-28 92%n=26 0.77 100% 2 Signal Studio claude-opus-4-7 cli 2026-05-28 96%n=26 0.81 100% 1 Signal Studio claude-sonnet-4-6 cli 2026-05-28 96%n=26 0.81 100% 1 AutoEdReport — 2026-05-27 92%n=26 0.77 100% 2 Signal Studio claude-opus-4-7 cli 2026-05-27 96%n=26 0.81 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 95%n=22 0.89 100% 1 Signal Studio claude-haiku-4-5 cli 2026-05-28 88%n=26 0.74 100% 3 Signal Studio claude-sonnet-4-6 cli 2026-05-28 96%n=26 0.81 100% 1 Signal Studio claude-opus-4-7 cli 2026-05-28 96%n=26 0.81 100% 1 AutoEdReport — 2026-05-27 92%n=26 0.77 100% 2 Signal Studio claude-opus-4-7 cli 2026-05-27 96%n=26 0.81 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 95%n=21 0.89 100% 1 Signal Studio claude-haiku-4-5 cli 2026-05-28 86%n=21 0.77 100% 3 Signal Studio claude-sonnet-4-6 cli 2026-05-28 100%n=21 1.00 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 95%n=21 0.89 100% 1 AutoEdReport — 2026-05-27 92%n=26 0.93 100% 2 Signal Studio claude-opus-4-7 cli 2026-05-27 100%n=26 1.00 100% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge exp-opus-4-7 claude-opus-4-7 cli 2026-05-31 75%n=24 0.65 67% 6 exp-opus-4-8 claude-opus-4-8 cli 2026-05-31 75%n=24 0.61 67% 6 exp-opus-4-7-vision claude-opus-4-7 cli 2026-05-31 71%n=24 0.43 67% 7 Signal Studio claude-opus-4-8 cli 2026-05-31 75%n=24 0.61 67% 6 Signal Studio claude-opus-4-7 cli 2026-05-28 75%n=24 0.47 67% 6 Signal Studio claude-sonnet-4-6 cli 2026-05-28 75%n=24 0.49 67% 6 Signal Studio claude-haiku-4-5 cli 2026-05-28 54%n=24 0.32 0% 11
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge AutoEdReport — 2026-05-24 — — 100% 0 AutoEdReport — 2026-05-24 — — 100% 0 AutoEdReport — 2026-05-24 — — 67% 0 Signal Studio claude-sonnet-4-6 2026-05-24 — — 0% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 — — 100% 0 Signal Studio claude-sonnet-4-6 cli 2026-05-28 — — 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 — — 50% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 — — 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 — — 100% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 96%n=25 1.00 100% 1 Signal Studio claude-haiku-4-5 cli 2026-05-28 88%n=25 0.91 100% 3 Signal Studio claude-sonnet-4-6 cli 2026-05-28 96%n=25 1.00 100% 1 Signal Studio claude-opus-4-7 cli 2026-05-28 96%n=25 1.00 100% 1 AutoEdReport — 2026-05-27 — — 67% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 96%n=25 1.00 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 96%n=25 1.00 100% 1 Signal Studio claude-haiku-4-5 cli 2026-05-28 92%n=25 0.99 100% 2 Signal Studio claude-sonnet-4-6 cli 2026-05-28 96%n=25 1.00 100% 1 Signal Studio claude-opus-4-7 cli 2026-05-28 96%n=25 1.00 100% 1 Signal Studio claude-opus-4-7 cli 2026-05-27 96%n=25 1.00 100% 1
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 92%n=25 0.99 100% 2 Signal Studio claude-haiku-4-5 cli 2026-05-28 80%n=25 0.80 67% 5 Signal Studio claude-sonnet-4-6 cli 2026-05-28 84%n=25 0.98 100% 4 Signal Studio claude-opus-4-7 cli 2026-05-28 92%n=25 0.99 100% 2 AutoEdReport — 2026-05-27 — — 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 92%n=25 0.99 100% 2
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge ss-experiment gemini-2.5-pro api 2026-06-01 88%n=41 0.69 67% 5 ss-experiment gemini-2.5-flash api 2026-06-01 80%n=41 0.55 67% 8 exp-opus-4-7 claude-opus-4-7 cli 2026-05-31 90%n=41 0.90 100% 4 exp-opus-4-8 claude-opus-4-8 cli 2026-05-31 90%n=41 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-31 90%n=41 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-31 88%n=41 0.89 100% 5 ss-experiment claude-opus-4-7 cli 2026-05-31 90%n=41 0.90 100% 4 Signal Studio claude-opus-4-8 cli 2026-05-31 90%n=41 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-29 90%n=41 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-29 90%n=41 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-29 90%n=41 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-29 90%n=41 0.90 100% 4 Signal Studio claude-opus-4-7 cli 2026-05-28 90%n=41 0.90 100% 4 Signal Studio claude-opus-4-7 cli 2026-05-28 90%n=41 0.90 100% 4 Signal Studio claude-haiku-4-5 cli 2026-05-28 78%n=41 0.48 67% 9 Signal Studio claude-sonnet-4-6 cli 2026-05-28 83%n=41 0.65 67% 7 Signal Studio claude-opus-4-7 cli 2026-05-27 — — 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 90%n=41 0.90 100% 4 AutoEdReport — 2026-05-27 — — 67% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge exp-opus-4-7 claude-opus-4-7 cli 2026-05-31 89%n=28 0.85 100% 3 exp-opus-4-8 claude-opus-4-8 cli 2026-05-31 86%n=28 0.77 100% 4 Signal Studio claude-opus-4-8 cli 2026-05-31 89%n=28 0.85 100% 3 Signal Studio claude-opus-4-7 cli 2026-05-28 86%n=28 0.84 100% 4 Signal Studio claude-opus-4-7 cli 2026-05-28 82%n=28 0.76 100% 5 Signal Studio claude-sonnet-4-6 cli 2026-05-28 86%n=28 0.84 100% 4 Signal Studio claude-haiku-4-5 cli 2026-05-28 69%n=26 0.46 33% 8 Signal Studio claude-opus-4-7 cli 2026-05-26 89%n=28 0.85 100% 3 AutoEdReport — 2026-05-26 89%n=28 0.89 100% 3 AutoEdReport — 2026-05-26 86%n=28 0.88 67% 4 AutoEdReport — 2026-05-26 57%n=28 0.72 33% 12 AutoEdReport — 2026-05-26 86%n=28 0.93 100% 4 AutoEdReport — 2026-05-26 86%n=28 0.88 100% 4 AutoEdReport — 2026-05-26 54%n=28 0.62 0% 13 AutoEdReport — 2026-05-26 71%n=28 0.72 100% 8 AutoEdReport — 2026-05-26 68%n=28 0.70 100% 9 AutoEdReport — 2026-05-26 89%n=28 0.95 100% 3 AutoEdReport — 2026-05-26 82%n=28 0.76 100% 5 AutoEdReport — 2026-05-26 43%n=28 0.52 0% 16 AutoEdReport — 2026-05-26 79%n=28 0.74 100% 6 Signal Studio claude-sonnet-4-6 2026-05-25 — — 100% 0 Signal Studio claude-opus-4-7 2026-05-25 — — 100% 0 Signal Studio claude-opus-4-7 2026-05-25 — — 100% 0 Signal Studio claude-sonnet-4-6 2026-05-25 — — 33% 0 Signal Studio claude-sonnet-4-6 2026-05-25 — — 100% 0 Signal Studio claude-opus-4-7 2026-05-25 86%n=28 0.84 100% 4 Signal Studio claude-sonnet-4-6 2026-05-25 82%n=28 0.87 100% 5 Signal Studio claude-haiku-4-5 2026-05-25 61%n=28 0.76 67% 11 Signal Studio claude-sonnet-4-6 2026-05-24 — — 0% 0 Signal Studio claude-sonnet-4-6 2026-05-24 71%n=28 0.71 33% 8
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge ss-experiment gemini-2.5-pro api 2026-06-01 83%n=36 0.72 50% 6 ss-experiment gemini-2.5-flash api 2026-06-01 86%n=36 0.89 100% 5 ss-experiment claude-opus-4-7 cli 2026-05-31 89%n=36 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-31 78%n=36 0.77 50% 8 ss-experiment claude-opus-4-7 cli 2026-05-31 89%n=36 0.90 100% 4 Signal Studio claude-opus-4-8 cli 2026-05-31 89%n=36 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-29 89%n=36 0.90 100% 4 ss-experiment claude-opus-4-7 cli 2026-05-29 92%n=36 0.91 100% 3 ss-experiment claude-opus-4-7 cli 2026-05-29 86%n=36 0.86 100% 5 ss-experiment claude-opus-4-7 cli 2026-05-29 83%n=36 0.88 100% 6 Signal Studio claude-opus-4-7 cli 2026-05-28 92%n=36 0.91 100% 3 Signal Studio claude-haiku-4-5 cli 2026-05-28 83%n=36 0.82 67% 6 Signal Studio claude-sonnet-4-6 cli 2026-05-28 83%n=36 0.85 100% 6 Signal Studio claude-opus-4-7 cli 2026-05-26 89%n=36 0.90 100% 4 AutoEdReport — 2026-05-26 78%n=36 0.77 100% 8 AutoEdReport — 2026-05-26 81%n=36 0.74 100% 7 AutoEdReport — 2026-05-26 81%n=36 0.83 100% 7 AutoEdReport — 2026-05-26 44%n=36 0.67 0% 20 AutoEdReport — 2026-05-26 67%n=36 0.66 100% 12 AutoEdReport — 2026-05-26 78%n=36 0.73 100% 8 AutoEdReport — 2026-05-26 83%n=36 0.88 100% 6 AutoEdReport — 2026-05-26 39%n=36 0.53 0% 22 AutoEdReport — 2026-05-26 78%n=36 0.72 67% 8 AutoEdReport — 2026-05-26 36%n=36 0.49 33% 23 AutoEdReport — 2026-05-26 63%n=30 0.65 — 11
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 87%n=39 0.90 100% 5 ss-experiment claude-opus-4-7 cli 2026-05-31 88%n=34 0.90 67% 4 ss-experiment claude-opus-4-7 cli 2026-05-31 87%n=39 0.90 100% 5 ss-experiment claude-opus-4-7 cli 2026-05-31 85%n=39 0.89 100% 6 ss-experiment claude-opus-4-7 cli 2026-05-29 85%n=39 0.89 100% 6 ss-experiment claude-opus-4-7 cli 2026-05-29 87%n=39 0.90 100% 5 ss-experiment claude-opus-4-7 cli 2026-05-29 85%n=39 0.89 100% 6 ss-experiment claude-opus-4-7 cli 2026-05-29 82%n=39 0.87 100% 7 Signal Studio claude-haiku-4-5 cli 2026-05-28 69%n=39 0.76 100% 12 Signal Studio claude-sonnet-4-6 cli 2026-05-28 79%n=39 0.83 100% 8 Signal Studio claude-opus-4-7 cli 2026-05-28 87%n=39 0.90 100% 5 Signal Studio claude-opus-4-7 cli 2026-05-26 85%n=39 0.89 100% 6 AutoEdReport — 2026-05-26 87%n=39 0.91 100% 5 AutoEdReport — 2026-05-26 87%n=39 0.91 100% 5 AutoEdReport — 2026-05-26 82%n=39 0.89 100% 7 AutoEdReport — 2026-05-26 79%n=39 0.85 100% 8 AutoEdReport — 2026-05-26 82%n=39 0.85 100% 7 AutoEdReport — 2026-05-26 51%n=39 0.73 67% 19 AutoEdReport — 2026-05-26 82%n=39 0.89 100% 7 AutoEdReport — 2026-05-26 46%n=39 0.65 67% 21 AutoEdReport — 2026-05-26 51%n=39 0.72 67% 19 AutoEdReport — 2026-05-26 48%n=21 0.12 — 11
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 83%n=30 0.89 100% 5 Signal Studio claude-opus-4-7 cli 2026-05-28 87%n=30 0.90 100% 4 AutoEdReport — 2026-05-28 — — 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 83%n=30 0.84 100% 5 Signal Studio claude-haiku-4-5 cli 2026-05-28 36%n=25 0.41 0% 16 Signal Studio claude-sonnet-4-6 cli 2026-05-28 77%n=30 0.87 100% 7
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 89%n=28 0.85 100% 3 Signal Studio claude-haiku-4-5 cli 2026-05-28 64%n=28 0.63 33% 10 Signal Studio claude-opus-4-7 cli 2026-05-28 86%n=28 0.84 100% 4 Signal Studio claude-sonnet-4-6 cli 2026-05-28 82%n=28 0.83 100% 5 AutoEdReport — 2026-05-27 82%n=28 0.88 67% 5 AutoEdReport — 2026-05-27 93%n=28 0.91 100% 2 AutoEdReport — 2026-05-27 86%n=28 0.88 67% 4 Signal Studio claude-opus-4-7 cli 2026-05-27 89%n=28 0.85 100% 3
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 93%n=28 0.91 100% 2 Signal Studio claude-sonnet-4-6 cli 2026-05-28 82%n=28 0.82 100% 5 Signal Studio claude-haiku-4-5 cli 2026-05-28 65%n=26 0.50 33% 9 Signal Studio claude-opus-4-7 cli 2026-05-28 79%n=28 0.69 100% 6 AutoEdReport — 2026-05-27 89%n=28 0.89 100% 3 AutoEdReport — 2026-05-27 86%n=28 0.88 100% 4 AutoEdReport — 2026-05-27 89%n=28 0.90 67% 3 Signal Studio claude-opus-4-7 cli 2026-05-27 89%n=28 0.85 100% 3
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge exp-opus-4-7 claude-opus-4-7 cli 2026-05-31 73%n=33 0.71 — 9 exp-opus-4-8 claude-opus-4-8 cli 2026-05-31 68%n=41 0.65 50% 13 AutoEdReport — 2026-05-28 — — 100% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge AutoEdReport — 2026-05-28 — — 0% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-opus-4-8 cli 2026-05-31 — — 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 — — 100% 0 Signal Studio claude-haiku-4-5 cli 2026-05-28 — — 0% 0 Signal Studio claude-sonnet-4-6 cli 2026-05-28 — — 100% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 — — 100% 0
Scored judge runs by source, measured against gold Source Model Date Indicator κw Gateway Diverge Signal Studio claude-haiku-4-5 cli 2026-05-28 — — 100% 0 Signal Studio claude-sonnet-4-6 cli 2026-05-28 — — 67% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 — — 67% 0 Signal Studio claude-opus-4-7 cli 2026-05-28 — — 67% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 — — 67% 0 Signal Studio claude-opus-4-7 cli 2026-05-27 — — 67% 0