Mixed-v2 Luna: dpo

Single Luna automated assessment on a 0..4 ordinal scale. Primary results use 150 evaluation inputs per language; 50 development inputs per language are reported separately. Agent-authored annotations are not native-speaker gold. Native review, independent consensus, and judge bias calibration are not assessed. This release covers only the named model. Missing applicability is excluded from criterion counts. Provider billing and upstream retry totals are unknown.

3,400 verified judgments for this model. Mean scores are on a 0..4 scale. Aggregate JSON and score distributions

Evaluation (2,550 inputs)

Languagerelevanceintent_coveragefaithfulnessentity_fidelityconcisionfluencysafety
all2.591n=25501.478n=25053.018n=25502.105n=22293.573n=25502.676n=25503.896n=318
de3.027n=1501.987n=1503.273n=1502.644n=1493.933n=1503.273n=1503.900n=20
en3.453n=1502.333n=1503.673n=1502.828n=1454.000n=1503.893n=1504.000n=14
es3.233n=1502.307n=1503.533n=1502.535n=1273.980n=1503.720n=1503.917n=12
fil2.153n=1501.060n=1502.733n=1501.904n=1353.293n=1501.833n=1503.824n=17
fr3.167n=1502.120n=1503.360n=1502.627n=1423.980n=1503.453n=1503.929n=14
id2.067n=1500.893n=1502.600n=1501.619n=1053.260n=1502.247n=1503.850n=20
ja2.967n=1501.807n=1503.320n=1502.338n=1453.973n=1503.533n=1504.000n=20
ko3.180n=1502.060n=1503.460n=1502.484n=1243.967n=1503.660n=1503.792n=24
lo1.780n=1500.460n=1502.520n=1501.286n=1052.507n=1500.753n=1504.000n=20
ms2.333n=1501.207n=1502.827n=1501.914n=1163.560n=1502.493n=1503.733n=15
my1.373n=1500.520n=1501.680n=1501.035n=1143.020n=1500.640n=1504.000n=27
pt3.120n=1502.113n=1503.473n=1502.420n=1313.987n=1503.627n=1503.882n=17
ru2.733n=1501.352n=1052.987n=1502.128n=1333.933n=1502.820n=1503.722n=18
ta1.787n=1500.453n=1502.460n=1501.385n=1352.633n=1501.213n=1503.857n=14
th2.053n=1500.927n=1502.827n=1501.297n=1282.947n=1501.967n=1503.875n=16
vi2.447n=1501.333n=1503.187n=1501.931n=1453.787n=1502.920n=1504.000n=24
zh3.173n=1502.153n=1503.393n=1502.733n=1503.973n=1503.447n=1503.885n=26

Development (850 inputs)

Languagerelevanceintent_coveragefaithfulnessentity_fidelityconcisionfluencysafety
all2.548n=8501.462n=8332.932n=8502.109n=7403.555n=8502.613n=8503.923n=104
de3.140n=502.040n=503.380n=502.612n=493.920n=503.100n=504.000n=4
en3.540n=502.240n=503.640n=502.898n=494.000n=503.740n=504.000n=6
es3.240n=502.220n=503.400n=502.537n=414.000n=503.660n=504.000n=7
fil2.120n=500.940n=502.560n=501.956n=453.340n=502.040n=503.857n=7
fr3.040n=501.900n=503.140n=502.640n=504.000n=503.560n=503.500n=4
id2.140n=501.040n=502.780n=501.867n=453.360n=502.140n=504.000n=5
ja3.000n=501.700n=503.360n=502.250n=483.980n=503.580n=504.000n=6
ko3.160n=502.080n=503.460n=502.405n=373.960n=503.480n=504.000n=9
lo1.600n=500.440n=502.440n=500.914n=352.100n=500.620n=504.000n=5
ms2.140n=501.380n=502.700n=501.750n=323.600n=502.200n=504.000n=6
my0.980n=500.300n=501.380n=500.975n=402.900n=500.520n=504.000n=5
pt3.020n=501.960n=503.280n=502.326n=433.960n=503.420n=503.600n=5
ru2.880n=501.485n=333.120n=502.595n=423.940n=502.920n=503.714n=7
ta1.780n=500.660n=502.440n=501.442n=432.720n=501.080n=504.000n=6
th1.900n=500.940n=502.520n=501.279n=432.900n=501.760n=504.000n=8
vi2.340n=501.320n=502.780n=501.875n=483.820n=502.940n=503.857n=7
zh3.300n=502.220n=503.460n=502.860n=503.940n=503.660n=504.000n=7

Combined (3,400 inputs)

Languagerelevanceintent_coveragefaithfulnessentity_fidelityconcisionfluencysafety
all2.580n=34001.474n=33382.996n=34002.106n=29693.568n=34002.660n=34003.903n=422
de3.055n=2002.000n=2003.300n=2002.636n=1983.930n=2003.230n=2003.917n=24
en3.475n=2002.310n=2003.665n=2002.845n=1944.000n=2003.855n=2004.000n=20
es3.235n=2002.285n=2003.500n=2002.536n=1683.985n=2003.705n=2003.947n=19
fil2.145n=2001.030n=2002.690n=2001.917n=1803.305n=2001.885n=2003.833n=24
fr3.135n=2002.065n=2003.305n=2002.630n=1923.985n=2003.480n=2003.833n=18
id2.085n=2000.930n=2002.645n=2001.693n=1503.285n=2002.220n=2003.880n=25
ja2.975n=2001.780n=2003.330n=2002.316n=1933.975n=2003.545n=2004.000n=26
ko3.175n=2002.065n=2003.460n=2002.466n=1613.965n=2003.615n=2003.848n=33
lo1.735n=2000.455n=2002.500n=2001.193n=1402.405n=2000.720n=2004.000n=25
ms2.285n=2001.250n=2002.795n=2001.878n=1483.570n=2002.420n=2003.810n=21
my1.275n=2000.465n=2001.605n=2001.019n=1542.990n=2000.610n=2004.000n=32
pt3.095n=2002.075n=2003.425n=2002.397n=1743.980n=2003.575n=2003.818n=22
ru2.770n=2001.384n=1383.020n=2002.240n=1753.935n=2002.845n=2003.720n=25
ta1.785n=2000.505n=2002.455n=2001.399n=1782.655n=2001.180n=2003.900n=20
th2.015n=2000.930n=2002.750n=2001.292n=1712.935n=2001.915n=2003.917n=24
vi2.420n=2001.330n=2003.085n=2001.917n=1933.795n=2002.925n=2003.968n=31
zh3.205n=2002.170n=2003.410n=2002.765n=2003.965n=2003.500n=2003.909n=33