Gemini 3.5 Flash-Lite

4 runs · 4 datasets · 1 model

slug: gemini-3-5-flash-lite

0.836

Best SPS · opinionsqa

Disaggregated subgroup scorecard. Each card below is one published run for this vendor; expand the question-type and demographic-subgroup sections to see the matrix beneath the headline SPS. Where coverage permits, 95% CI bands accompany the point estimate.

What each cell measures

Each cell compares this vendor's synthetic responses against real human respondents in that demographic subgroup — published survey ground truth, not a model's guess about the subgroup. Higher = closer to how that real subgroup actually answered.

Demographic conditioning here is not stereotyping: every conditioned score is checked against what real subgroup members said, not against assumptions about them — and where the vendor's output diverges from the real subgroup, the score drops.

globalopinionqa — ground truth: Durmus et al. 2023, Anthropic — llm_global_opinions
gss — ground truth: NORC at the University of Chicago — General Social Survey
opinionsqa — ground truth: Santurkar et al., ICML 2023 — Whose Opinions Do LLMs Reflect? (derived from Pew American Trends Panel)
subpop — ground truth: Suh et al., ACL 2025 — SubPOP: Subpopulation-Level Opinion Prediction

Low-n cells: cells with n < 30 are shown muted and tagged low n — suggestive only instead of color-graded; they are reported for transparency, not as findings.

CIs: "no CI — single run" marks point estimates from a single run with no confidence interval yet; treat the uncertainty as unknown, never zero.

Columns: Score = distributional parity (p_dist) restricted to the subgroup · p_cond = conditioning strength vs the unconditioned baseline · N = questions answered under that conditioning · Cov. = coverage dot derived from N (green = high, ≥100 · amber = medium, 50–99 · red = low, <50). Topic tables: SPS = Survey Parity Score for the topic · p_dist = 1 − mean(JSD) · p_rank = (1 + mean(τ)) / 2 · p_refuse = 1 − mean(|R_model − R_human|).

Multiple comparisons: with this many subgroup cells, a few extreme cells are expected by chance alone — read patterns across a dimension, not single cells.

globalopinionqa raw

raw--gemini-3.5-flash-lite--tdefault--tplcurrent--8aa5510f

0.770 ± 0.033

SPS · 95% CI [0.735, 0.802] · n = 100

Question-type breakdown (6 topics)

Topic	SPS	p_dist	p_rank	p_refuse	N
Economy & Work low n — suggestive only	0.941	0.912	0.969	0.949	3
Health & Science low n — suggestive only	0.887	0.775	1.000	1.000	1
General Attitudes low n — suggestive only	0.772	0.794	0.750	1.000	4
International Relations & Security	0.685	0.679	0.690	0.987	60
Trust & Wellbeing low n — suggestive only	0.625	0.516	0.734	1.000	2
Politics & Governance	0.610	0.595	0.625	0.903	30

Topics with N < 10 are muted and tagged "low n — suggestive only": too few questions for a stable 3-decimal score.

Not yet measured

This vendor has no demographic-conditioned runs for age, geography, education, or any other subgroup dimension on globalopinionqa. No cells are fabricated — scores appear here only when a conditioned run actually measured them.

How to submit a demographic-conditioned run →

gss raw

raw--gemini-3.5-flash-lite--tdefault--tplcurrent--2e530450

0.807 ± 0.022

SPS · 95% CI [0.784, 0.828] · n = 75

Question-type breakdown (7 topics)

Topic	SPS	p_dist	p_rank	p_refuse	N
Health & Science low n — suggestive only	0.883	0.859	0.908	0.967	2
Social Values & Religion	0.790	0.743	0.838	0.977	14
International Relations & Security	0.725	0.706	0.745	0.974	33
Economy & Work low n — suggestive only	0.721	0.783	0.658	0.981	4
Politics & Governance low n — suggestive only	0.690	0.652	0.729	0.975	9
General Attitudes	0.660	0.627	0.694	0.960	10
Trust & Wellbeing low n — suggestive only	0.639	0.645	0.634	0.951	3

Topics with N < 10 are muted and tagged "low n — suggestive only": too few questions for a stable 3-decimal score.

Not yet measured

This vendor has no demographic-conditioned runs for age, geography, education, or any other subgroup dimension on gss. No cells are fabricated — scores appear here only when a conditioned run actually measured them.

How to submit a demographic-conditioned run →

opinionsqa raw

raw--gemini-3.5-flash-lite--tdefault--tplcurrent--dad20c6a

0.836 ± 0.018

SPS · 95% CI [0.818, 0.853] · n = 100

Question-type breakdown (9 topics)

Topic	SPS	p_dist	p_rank	p_refuse	N
Media & Information low n — suggestive only	0.892	0.877	0.908	0.996	2
Health & Science low n — suggestive only	0.806	0.758	0.854	0.997	1
Economy & Work	0.801	0.764	0.838	0.995	21
Technology & Digital Life low n — suggestive only	0.786	0.756	0.816	0.995	4
Trust & Wellbeing low n — suggestive only	0.776	0.753	0.799	0.994	1
International Relations & Security	0.768	0.751	0.785	0.994	20
Social Values & Religion	0.755	0.728	0.781	0.993	15
Politics & Governance low n — suggestive only	0.746	0.675	0.816	0.996	1
General Attitudes	0.713	0.686	0.740	0.994	35

Topics with N < 10 are muted and tagged "low n — suggestive only": too few questions for a stable 3-decimal score.

Not yet measured

This vendor has no demographic-conditioned runs for age, geography, education, or any other subgroup dimension on opinionsqa. No cells are fabricated — scores appear here only when a conditioned run actually measured them.

How to submit a demographic-conditioned run →

subpop raw

raw--gemini-3.5-flash-lite--tdefault--tplcurrent--30a43ab6

0.801 ± 0.020

SPS · 95% CI [0.781, 0.821] · n = 100

Question-type breakdown (9 topics)

Topic	SPS	p_dist	p_rank	p_refuse	N
Identity & Demographics low n — suggestive only	0.846	0.838	0.854	0.995	1
General Attitudes	0.794	0.764	0.823	0.936	17
Politics & Governance low n — suggestive only	0.760	0.737	0.784	0.981	4
Health & Science low n — suggestive only	0.760	0.710	0.810	0.994	6
Technology & Digital Life	0.712	0.662	0.763	0.993	16
Economy & Work	0.698	0.647	0.748	0.992	13
Social Values & Religion	0.679	0.626	0.731	0.986	27
International Relations & Security	0.658	0.609	0.708	0.991	15
Media & Information low n — suggestive only	0.466	0.475	0.457	0.978	1

Topics with N < 10 are muted and tagged "low n — suggestive only": too few questions for a stable 3-decimal score.

Not yet measured

This vendor has no demographic-conditioned runs for age, geography, education, or any other subgroup dimension on subpop. No cells are fabricated — scores appear here only when a conditioned run actually measured them.

How to submit a demographic-conditioned run →

No demographic conditioning data has been published for this vendor yet. The question-type matrix above shows topic-level parity; subgroup rows fill in once Althing-style conditioned runs land.

← Back to leaderboard