Skip to content
EdReports

Evaluation runs

Every scored judge run, treated as a rater and measured against gold. Track the model bake-off and Signal Studio vs AutoEdReport across the canonical curricula.

276 scored runs across 28 curricula

Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
ss-experimentgemini-2.5-proapi2026-06-0185%n=340.56100%5
ss-experimentgemini-2.5-flashapi2026-06-0176%n=340.5467%8
ss-experimentclaude-sonnet-4-6cli2026-06-0182%n=340.67100%6
ss-experimentclaude-haiku-4-5cli2026-06-0162%n=340.32100%13
timing-visionclaude-opus-4-7api2026-05-31100%n=61.00100%0
Signal Studioclaude-opus-4-8cli2026-05-3182%n=340.49100%6
vision-imgclaude-opus-4-7api2026-05-3182%n=340.72100%6
vision-baseclaude-opus-4-7api2026-05-3182%n=340.72100%6
Signal Studioclaude-opus-4-7cli2026-05-2882%n=340.67100%6
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%n=291.00100%0
Signal Studioclaude-opus-4-7cli2026-05-28100%n=291.00100%0
Signal Studioclaude-opus-4-7cli2026-05-28100%n=291.00100%0
Signal Studioclaude-sonnet-4-6cli2026-05-28100%n=291.00100%0
Signal Studioclaude-haiku-4-5cli2026-05-2893%n=290.94100%2
Signal Studioclaude-opus-4-7cli2026-05-26100%n=291.00100%0
AutoEdReport2026-05-26100%n=291.00100%0
AutoEdReport2026-05-26100%n=291.00100%0
AutoEdReport2026-05-26100%n=291.00100%0
AutoEdReport2026-05-2697%n=290.97100%1
AutoEdReport2026-05-2631%n=290.270%20
AutoEdReport2026-05-2648%n=290.280%15
AutoEdReport2026-05-2697%n=290.97100%1
Signal Studioclaude-sonnet-4-62026-05-25100%n=291.00100%0
Signal Studioclaude-sonnet-4-62026-05-25100%n=291.00100%0
Signal Studioclaude-opus-4-72026-05-25100%n=291.00100%0
Signal Studioclaude-opus-4-72026-05-25100%n=291.00100%0
Signal Studioclaude-sonnet-4-62026-05-25100%n=291.00100%0
Signal Studioclaude-opus-4-72026-05-25100%n=291.00100%0
Signal Studioclaude-sonnet-4-62026-05-25100%n=291.00100%0
Signal Studioclaude-haiku-4-52026-05-25100%n=291.00100%0
Signal Studioclaude-sonnet-4-62026-05-2490%n=290.80100%3
Signal Studioclaude-sonnet-4-62026-05-24100%n=291.00100%0
Signal Studioclaude-sonnet-4-62026-05-2497%n=290.86100%1
Signal Studioclaude-sonnet-4-62026-05-24100%n=291.00100%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%n=171.00100%0
Signal Studioclaude-haiku-4-5cli2026-05-2871%n=170.0050%5
Signal Studioclaude-sonnet-4-6cli2026-05-28100%n=171.00100%0
Signal Studioclaude-opus-4-7cli2026-05-28100%n=171.00100%0
Signal Studioclaude-opus-4-7cli2026-05-27100%n=171.00100%0
AutoEdReport2026-05-2794%n=170.00100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%n=191.00100%0
Signal Studioclaude-haiku-4-5cli2026-05-2884%n=190.4967%3
Signal Studioclaude-opus-4-7cli2026-05-28100%n=191.00100%0
Signal Studioclaude-sonnet-4-6cli2026-05-28100%n=191.00100%0
AutoEdReport2026-05-2784%n=190.73100%3
Signal Studioclaude-opus-4-7cli2026-05-2795%n=190.89100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%n=171.00100%0
Signal Studioclaude-haiku-4-5cli2026-05-2859%n=170.0050%7
Signal Studioclaude-sonnet-4-6cli2026-05-2888%n=170.00100%2
Signal Studioclaude-opus-4-7cli2026-05-28100%n=171.00100%0
AutoEdReport2026-05-2788%n=170.00100%2
Signal Studioclaude-opus-4-7cli2026-05-27100%n=171.00100%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%n=161.00100%0
Signal Studioclaude-haiku-4-5cli2026-05-2856%n=160.0050%7
Signal Studioclaude-sonnet-4-6cli2026-05-28100%n=161.00100%0
Signal Studioclaude-opus-4-7cli2026-05-28100%n=161.00100%0
AutoEdReport2026-05-27100%n=161.00100%0
Signal Studioclaude-opus-4-7cli2026-05-27100%n=161.00100%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-31100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-31100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-3195%n=190.89100%1
ss-experimentclaude-opus-4-7cli2026-05-31100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-31100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-31100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-29100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-29100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-29100%n=191.00100%0
ss-experimentclaude-opus-4-7cli2026-05-2995%n=190.89100%1
ss-experimentclaude-opus-4-7cli2026-05-29100%n=191.00100%0
Signal Studioclaude-opus-4-7cli2026-05-28100%n=191.00100%0
Signal Studioclaude-haiku-4-5cli2026-05-2884%n=190.5867%3
Signal Studioclaude-sonnet-4-6cli2026-05-28100%n=191.00100%0
Signal Studioclaude-opus-4-7cli2026-05-28100%n=191.00100%0
AutoEdReport2026-05-2784%n=190.73100%3
Signal Studioclaude-opus-4-7cli2026-05-2795%n=190.89100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3197%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-3197%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-3197%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-3197%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-3197%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-2997%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-2997%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-2997%n=300.90100%1
ss-experimentclaude-opus-4-7cli2026-05-2997%n=300.90100%1
Signal Studioclaude-opus-4-7cli2026-05-2897%n=300.90100%1
Signal Studioclaude-haiku-4-5cli2026-05-2879%n=290.73100%6
Signal Studioclaude-sonnet-4-6cli2026-05-2897%n=300.90100%1
Signal Studioclaude-opus-4-7cli2026-05-2897%n=300.90100%1
Signal Studioclaude-opus-4-7cli2026-05-2697%n=300.90100%1
AutoEdReport2026-05-2697%n=300.97100%1
Signal Studioclaude-opus-4-72026-05-2597%n=300.90100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3195%n=220.89100%1
Signal Studioclaude-haiku-4-5cli2026-05-2892%n=260.77100%2
Signal Studioclaude-opus-4-7cli2026-05-2896%n=260.81100%1
Signal Studioclaude-sonnet-4-6cli2026-05-2896%n=260.81100%1
AutoEdReport2026-05-2792%n=260.77100%2
Signal Studioclaude-opus-4-7cli2026-05-2796%n=260.81100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3195%n=220.89100%1
Signal Studioclaude-haiku-4-5cli2026-05-2888%n=260.74100%3
Signal Studioclaude-sonnet-4-6cli2026-05-2896%n=260.81100%1
Signal Studioclaude-opus-4-7cli2026-05-2896%n=260.81100%1
AutoEdReport2026-05-2792%n=260.77100%2
Signal Studioclaude-opus-4-7cli2026-05-2796%n=260.81100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3195%n=210.89100%1
Signal Studioclaude-haiku-4-5cli2026-05-2886%n=210.77100%3
Signal Studioclaude-sonnet-4-6cli2026-05-28100%n=211.00100%0
Signal Studioclaude-opus-4-7cli2026-05-2895%n=210.89100%1
AutoEdReport2026-05-2792%n=260.93100%2
Signal Studioclaude-opus-4-7cli2026-05-27100%n=261.00100%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
exp-opus-4-7claude-opus-4-7cli2026-05-3175%n=240.6567%6
exp-opus-4-8claude-opus-4-8cli2026-05-3175%n=240.6167%6
exp-opus-4-7-visionclaude-opus-4-7cli2026-05-3171%n=240.4367%7
Signal Studioclaude-opus-4-8cli2026-05-3175%n=240.6167%6
Signal Studioclaude-opus-4-7cli2026-05-2875%n=240.4767%6
Signal Studioclaude-sonnet-4-6cli2026-05-2875%n=240.4967%6
Signal Studioclaude-haiku-4-5cli2026-05-2854%n=240.320%11
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
AutoEdReport2026-05-24100%0
AutoEdReport2026-05-24100%0
AutoEdReport2026-05-2467%0
Signal Studioclaude-sonnet-4-62026-05-240%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%0
Signal Studioclaude-sonnet-4-6cli2026-05-28100%0
Signal Studioclaude-haiku-4-5cli2026-05-2850%0
Signal Studioclaude-opus-4-7cli2026-05-28100%0
Signal Studioclaude-opus-4-7cli2026-05-27100%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3196%n=251.00100%1
Signal Studioclaude-haiku-4-5cli2026-05-2888%n=250.91100%3
Signal Studioclaude-sonnet-4-6cli2026-05-2896%n=251.00100%1
Signal Studioclaude-opus-4-7cli2026-05-2896%n=251.00100%1
AutoEdReport2026-05-2767%0
Signal Studioclaude-opus-4-7cli2026-05-2796%n=251.00100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3196%n=251.00100%1
Signal Studioclaude-haiku-4-5cli2026-05-2892%n=250.99100%2
Signal Studioclaude-sonnet-4-6cli2026-05-2896%n=251.00100%1
Signal Studioclaude-opus-4-7cli2026-05-2896%n=251.00100%1
Signal Studioclaude-opus-4-7cli2026-05-2796%n=251.00100%1
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3192%n=250.99100%2
Signal Studioclaude-haiku-4-5cli2026-05-2880%n=250.8067%5
Signal Studioclaude-sonnet-4-6cli2026-05-2884%n=250.98100%4
Signal Studioclaude-opus-4-7cli2026-05-2892%n=250.99100%2
AutoEdReport2026-05-27100%0
Signal Studioclaude-opus-4-7cli2026-05-2792%n=250.99100%2
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
ss-experimentgemini-2.5-proapi2026-06-0188%n=410.6967%5
ss-experimentgemini-2.5-flashapi2026-06-0180%n=410.5567%8
exp-opus-4-7claude-opus-4-7cli2026-05-3190%n=410.90100%4
exp-opus-4-8claude-opus-4-8cli2026-05-3190%n=410.90100%4
ss-experimentclaude-opus-4-7cli2026-05-3190%n=410.90100%4
ss-experimentclaude-opus-4-7cli2026-05-3188%n=410.89100%5
ss-experimentclaude-opus-4-7cli2026-05-3190%n=410.90100%4
Signal Studioclaude-opus-4-8cli2026-05-3190%n=410.90100%4
ss-experimentclaude-opus-4-7cli2026-05-2990%n=410.90100%4
ss-experimentclaude-opus-4-7cli2026-05-2990%n=410.90100%4
ss-experimentclaude-opus-4-7cli2026-05-2990%n=410.90100%4
ss-experimentclaude-opus-4-7cli2026-05-2990%n=410.90100%4
Signal Studioclaude-opus-4-7cli2026-05-2890%n=410.90100%4
Signal Studioclaude-opus-4-7cli2026-05-2890%n=410.90100%4
Signal Studioclaude-haiku-4-5cli2026-05-2878%n=410.4867%9
Signal Studioclaude-sonnet-4-6cli2026-05-2883%n=410.6567%7
Signal Studioclaude-opus-4-7cli2026-05-27100%0
Signal Studioclaude-opus-4-7cli2026-05-2790%n=410.90100%4
AutoEdReport2026-05-2767%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
exp-opus-4-7claude-opus-4-7cli2026-05-3189%n=280.85100%3
exp-opus-4-8claude-opus-4-8cli2026-05-3186%n=280.77100%4
Signal Studioclaude-opus-4-8cli2026-05-3189%n=280.85100%3
Signal Studioclaude-opus-4-7cli2026-05-2886%n=280.84100%4
Signal Studioclaude-opus-4-7cli2026-05-2882%n=280.76100%5
Signal Studioclaude-sonnet-4-6cli2026-05-2886%n=280.84100%4
Signal Studioclaude-haiku-4-5cli2026-05-2869%n=260.4633%8
Signal Studioclaude-opus-4-7cli2026-05-2689%n=280.85100%3
AutoEdReport2026-05-2689%n=280.89100%3
AutoEdReport2026-05-2686%n=280.8867%4
AutoEdReport2026-05-2657%n=280.7233%12
AutoEdReport2026-05-2686%n=280.93100%4
AutoEdReport2026-05-2686%n=280.88100%4
AutoEdReport2026-05-2654%n=280.620%13
AutoEdReport2026-05-2671%n=280.72100%8
AutoEdReport2026-05-2668%n=280.70100%9
AutoEdReport2026-05-2689%n=280.95100%3
AutoEdReport2026-05-2682%n=280.76100%5
AutoEdReport2026-05-2643%n=280.520%16
AutoEdReport2026-05-2679%n=280.74100%6
Signal Studioclaude-sonnet-4-62026-05-25100%0
Signal Studioclaude-opus-4-72026-05-25100%0
Signal Studioclaude-opus-4-72026-05-25100%0
Signal Studioclaude-sonnet-4-62026-05-2533%0
Signal Studioclaude-sonnet-4-62026-05-25100%0
Signal Studioclaude-opus-4-72026-05-2586%n=280.84100%4
Signal Studioclaude-sonnet-4-62026-05-2582%n=280.87100%5
Signal Studioclaude-haiku-4-52026-05-2561%n=280.7667%11
Signal Studioclaude-sonnet-4-62026-05-240%0
Signal Studioclaude-sonnet-4-62026-05-2471%n=280.7133%8
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
ss-experimentgemini-2.5-proapi2026-06-0183%n=360.7250%6
ss-experimentgemini-2.5-flashapi2026-06-0186%n=360.89100%5
ss-experimentclaude-opus-4-7cli2026-05-3189%n=360.90100%4
ss-experimentclaude-opus-4-7cli2026-05-3178%n=360.7750%8
ss-experimentclaude-opus-4-7cli2026-05-3189%n=360.90100%4
Signal Studioclaude-opus-4-8cli2026-05-3189%n=360.90100%4
ss-experimentclaude-opus-4-7cli2026-05-2989%n=360.90100%4
ss-experimentclaude-opus-4-7cli2026-05-2992%n=360.91100%3
ss-experimentclaude-opus-4-7cli2026-05-2986%n=360.86100%5
ss-experimentclaude-opus-4-7cli2026-05-2983%n=360.88100%6
Signal Studioclaude-opus-4-7cli2026-05-2892%n=360.91100%3
Signal Studioclaude-haiku-4-5cli2026-05-2883%n=360.8267%6
Signal Studioclaude-sonnet-4-6cli2026-05-2883%n=360.85100%6
Signal Studioclaude-opus-4-7cli2026-05-2689%n=360.90100%4
AutoEdReport2026-05-2678%n=360.77100%8
AutoEdReport2026-05-2681%n=360.74100%7
AutoEdReport2026-05-2681%n=360.83100%7
AutoEdReport2026-05-2644%n=360.670%20
AutoEdReport2026-05-2667%n=360.66100%12
AutoEdReport2026-05-2678%n=360.73100%8
AutoEdReport2026-05-2683%n=360.88100%6
AutoEdReport2026-05-2639%n=360.530%22
AutoEdReport2026-05-2678%n=360.7267%8
AutoEdReport2026-05-2636%n=360.4933%23
AutoEdReport2026-05-2663%n=300.6511
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3187%n=390.90100%5
ss-experimentclaude-opus-4-7cli2026-05-3188%n=340.9067%4
ss-experimentclaude-opus-4-7cli2026-05-3187%n=390.90100%5
ss-experimentclaude-opus-4-7cli2026-05-3185%n=390.89100%6
ss-experimentclaude-opus-4-7cli2026-05-2985%n=390.89100%6
ss-experimentclaude-opus-4-7cli2026-05-2987%n=390.90100%5
ss-experimentclaude-opus-4-7cli2026-05-2985%n=390.89100%6
ss-experimentclaude-opus-4-7cli2026-05-2982%n=390.87100%7
Signal Studioclaude-haiku-4-5cli2026-05-2869%n=390.76100%12
Signal Studioclaude-sonnet-4-6cli2026-05-2879%n=390.83100%8
Signal Studioclaude-opus-4-7cli2026-05-2887%n=390.90100%5
Signal Studioclaude-opus-4-7cli2026-05-2685%n=390.89100%6
AutoEdReport2026-05-2687%n=390.91100%5
AutoEdReport2026-05-2687%n=390.91100%5
AutoEdReport2026-05-2682%n=390.89100%7
AutoEdReport2026-05-2679%n=390.85100%8
AutoEdReport2026-05-2682%n=390.85100%7
AutoEdReport2026-05-2651%n=390.7367%19
AutoEdReport2026-05-2682%n=390.89100%7
AutoEdReport2026-05-2646%n=390.6567%21
AutoEdReport2026-05-2651%n=390.7267%19
AutoEdReport2026-05-2648%n=210.1211
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3183%n=300.89100%5
Signal Studioclaude-opus-4-7cli2026-05-2887%n=300.90100%4
AutoEdReport2026-05-28100%0
Signal Studioclaude-opus-4-7cli2026-05-2883%n=300.84100%5
Signal Studioclaude-haiku-4-5cli2026-05-2836%n=250.410%16
Signal Studioclaude-sonnet-4-6cli2026-05-2877%n=300.87100%7
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3189%n=280.85100%3
Signal Studioclaude-haiku-4-5cli2026-05-2864%n=280.6333%10
Signal Studioclaude-opus-4-7cli2026-05-2886%n=280.84100%4
Signal Studioclaude-sonnet-4-6cli2026-05-2882%n=280.83100%5
AutoEdReport2026-05-2782%n=280.8867%5
AutoEdReport2026-05-2793%n=280.91100%2
AutoEdReport2026-05-2786%n=280.8867%4
Signal Studioclaude-opus-4-7cli2026-05-2789%n=280.85100%3
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-3193%n=280.91100%2
Signal Studioclaude-sonnet-4-6cli2026-05-2882%n=280.82100%5
Signal Studioclaude-haiku-4-5cli2026-05-2865%n=260.5033%9
Signal Studioclaude-opus-4-7cli2026-05-2879%n=280.69100%6
AutoEdReport2026-05-2789%n=280.89100%3
AutoEdReport2026-05-2786%n=280.88100%4
AutoEdReport2026-05-2789%n=280.9067%3
Signal Studioclaude-opus-4-7cli2026-05-2789%n=280.85100%3
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
exp-opus-4-7claude-opus-4-7cli2026-05-3173%n=330.719
exp-opus-4-8claude-opus-4-8cli2026-05-3168%n=410.6550%13
AutoEdReport2026-05-28100%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
AutoEdReport2026-05-280%0
ELA

EngageNY ELA

5 runs
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-opus-4-8cli2026-05-31100%0
Signal Studioclaude-opus-4-7cli2026-05-28100%0
Signal Studioclaude-haiku-4-5cli2026-05-280%0
Signal Studioclaude-sonnet-4-6cli2026-05-28100%0
Signal Studioclaude-opus-4-7cli2026-05-27100%0
Scored judge runs by source, measured against gold
SourceModelDateIndicatorκwGatewayDiverge
Signal Studioclaude-haiku-4-5cli2026-05-28100%0
Signal Studioclaude-sonnet-4-6cli2026-05-2867%0
Signal Studioclaude-opus-4-7cli2026-05-2867%0
Signal Studioclaude-opus-4-7cli2026-05-2867%0
Signal Studioclaude-opus-4-7cli2026-05-2767%0
Signal Studioclaude-opus-4-7cli2026-05-2767%0