Gemini 3.8 Flash: score and benchmark evidence
Historical release · · Google
104.0 points · Rank 9 among 19 ranked models · Stability range 78–131.3 points.
Rank follows the point estimate; it does not establish statistically significant superiority. The scale is anchored at mean 100, SD 15 in the frozen calibration cohort, not human IQ or a percentage.
One capability score combines evidence across seven equally weighted domains. Related benchmark variants share a family budget, and influence from one evaluator is capped.
Release tli-2026-v5.0-2026-09-06. Methodology and limitations for this release.
This page preserves historical evidence. See the current model record.
Capability domains
Domain scores use calibration-cohort standard-deviation units. Unmeasured domains are shown as unavailable, not zero.
| Domain | Score (SD units) | Observed benchmarks |
|---|---|---|
| Knowledge & reasoning | 0.34 | 6 |
| Coding & software engineering | 0.04 | 5 |
| Agentic tool & computer use | 0.25 | 2 |
| Professional & real-world work | 0.08 | 10 |
| Multimodal & vision | -0.15 | 3 |
| Cybersecurity | Unavailable | 0 |
| Preference & communication | Unavailable | 0 |
Published benchmark evidence
26 results contribute to this score, from 52 collected results and 26 contributing benchmark families. Counts are not independent sample sizes.
Only results with a matched source record, metric and documented configuration contribute. Developer-reported results are eligible when their configuration is clear. Unresolved records remain in the database with their exclusion reason. Archived reviews and live source checks are labelled separately.
The interval reflects limited evidence, missing domains, correlated benchmark families and frozen calibration uncertainty under the scoring assumptions. It is not a promise that future scores will fall inside it 95% of the time. New conflicting evidence can widen it.
Reporting sensitivity: ±4.7 points. Sensitivity to selectively published results. This stress range is not a confidence interval or a probability statement.
GDPval-AA v2 (Elo): 1545 native-points
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
Terminal-Bench 2.1: 89.4%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
DeepSWE v1.1: 73.7%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
LMArena (Chatbot Arena) — Text: 1494 native-points
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
AA Intelligence Index v4.1.1 (score): 59 native-points
Not used in the score. Source category not recorded. Review status: excluded.
Held-out composite, never a scoring input
GPQA Diamond: 95.3%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Gemini 3.8 Flash (high); AA published evaluation harness
Metric: accuracy, AA
Original source · Retrieved 2026-09-06
Current first-party evaluator data matches the recorded score at its published rounding. Exact pinned configuration selected; later value changes are not silently substituted.
AA-LCR: 82%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
Finance Agent v2 — Vals: 61.44%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: Finance Agent (v2) overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
Excel Modeling Benchmark — Vals: 72.2%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: EMB overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
SciCode: 54.4%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
Humanity's Last Exam (no tools): 47.8%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Gemini 3.8 Flash (high); AA published evaluation harness
Metric: accuracy, no tools, AA
Original source · Retrieved 2026-09-06
Current first-party evaluator data matches the recorded score at its published rounding. Exact pinned configuration selected; later value changes are not silently substituted.
Legal Research Bench — Vals: 38.94%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: Legal Research Bench overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
Harvey LAB (Vals): 10%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: Harvey's Legal Agent Benchmark overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
AA-Briefcase Overall Elo: 1213 native-points
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
MedScribe — Vals: 84.5%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: MedScribe overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
SWE-bench Verified — Vals: 80%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
LegalBench — Vals: 86.99%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: LegalBench overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
TaxEval v2 — Vals: 74.45%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: TaxEval v2 overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
Vals Index: 62.25%
Not used in the score. Source category not recorded. Review status: excluded.
Held-out composite, never a scoring input
Public Benefits Bench — Vals: 65.29%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: Public Benefits Bench v1.1 overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
MedCode — Vals: 48.13%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: MedCode overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
AA-Omniscience: 29.6 native-points
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Gemini 3.8 Flash (high); AA published evaluation harness
Metric: AA Omniscience index -100 to 100
Original source · Retrieved 2026-09-06
Current first-party evaluator data matches the recorded score at its published rounding. Exact pinned configuration selected; later value changes are not silently substituted.
MMLU Pro — Vals: 90.22%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: MMLU Pro overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
LiveBench: 75.83%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
MortgageTax — Vals: 65.34%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: MortgageTax overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
SAGE — Vals: 35.06%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: SAGE overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
Code Migration †: 36.55%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: Code Migration overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
LiveCodeBench: 89.48%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: LiveCodeBench overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
ProgramBench v1 (Raw Pass Rate) — Vals: 71.899%
Contributes to the score. Source category: independent. Review status: reviewed-archive.
Configuration: google/gemini-3.8-flash · reasoning_effort=high
Metric: Raw Pass Rate: mean percent of hidden behavioral tests passed per task. Not Fully Resolved (all tests pass) and not Almost Resolved (≥95% of tests pass).
Original source · Retrieved 2026-09-05
Matched exact metric/value and the archived benchmark-specific configuration review. Unknown native effort stays explicitly unknown.
SkillsBench †: 57.98%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
CursorBench v3.2: 69.2%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
SWE-bench Pro: 61.6%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
CharXiv (with CI / RQ): 86.2%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
MMMU-Pro: 89.08%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
FrontierCode v1.1 Extended: 56.3%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
LVBench (with Memory): 87.8%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
MathArena — ArXivMath Jun 2026: 68.03%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
MathArena — BrokenArXiv Jun 2026: 20.83%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
FrontierCode v1.1 Main — Anthropic H2H: 43.6%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
τ³-Banking: 44.9%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Gemini 3.8 Flash (high); AA published evaluation harness
Metric: strict task success, AA
Original source · Retrieved 2026-09-06
Current first-party evaluator data matches the recorded score at its published rounding. Exact pinned configuration selected; later value changes are not silently substituted.
CritPt: 18.3%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Gemini 3.8 Flash (high); AA published evaluation harness
Metric: accuracy, AA official grading
Original source · Retrieved 2026-09-06
Current first-party evaluator data matches the recorded score at its published rounding. Exact pinned configuration selected; later value changes are not silently substituted.
Terminal-Bench 4.0: 19.1%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
ProofBench v1.1 — Vals: 48%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: ProofBench v1.1 overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
Vibe Code Bench v1.1 — Vals: 78.65%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: google_gemini-3.8-flash; Vals high effort
Metric: Vibe Code Bench v1.1 overall percentage; Vals benchmark-specific harness
Original source · Retrieved 2026-09-06
Exact model and benchmark row; current source value agrees at sheet rounding. Uses published default parameters; source warns individual benchmarks may override them. Known mismatched rows/harnesses are not imported.
MineBench: 1917 native-points
Not used in the score. Source category not recorded. Review status: excluded.
Unresolved Pro/base model identities in source catalog; exclude until matched.
HealthBench Professional (length-adjusted): 52.1%
Not used in the score. Source category not recorded. Review status: unresolved.
Source selects the maximum score across efforts; exact highest-effort result is not established.
BioMysteryBench v1 — Vals: 62.222%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
Tax Agent Bench v1 — Vals: 66.771%
Not used in the score. Source category not recorded. Review status: unresolved.
Inherited cell lacks a matched, reviewed source/configuration record.
RuneBench (30m, mean ln(1 + XP/min)): 5.962732 native-points
Contributes to the score. Source category: independent. Review status: reviewed-archive.
Configuration: gemini38flash
Metric: Arithmetic mean of ln(1 + peakXpRate) over all 16 skills, exactly as the owner’s Heatmap.js default ranking. Data rates are already normalized; do not divide by 200 again.
Original source · Retrieved 2026-09-06
Matched exact metric/value and the archived benchmark-specific configuration review. Unknown native effort stays explicitly unknown.
VoxelBench (text-to-voxel, Glicko-2): 1811 native-points
Contributes to the score. Source category: independent. Review status: reviewed-archive.
Configuration: Gemini 3.8 Flash
Metric: Glicko-2 rating from the text-to-voxel board; not win percentage, deviation, image-input mode or MineBench Elo.
Original source · Retrieved 2026-09-06
Matched exact metric/value and the archived benchmark-specific configuration review. Unknown native effort stays explicitly unknown.
Roboflow Vision Evals (six-task mean, high tier): 86.516667%
Contributes to the score. Source category: independent. Review status: reviewed-archive.
Configuration: Require a complete six-task high-tier panel, the highest published benchmark tier. Retain high-tier scores even when lower than low-tier. Do not substitute the best of three repetitions: use the published mean. Each panel enters the fit once; do not also fit its components.
Metric: Arithmetic mean of six published high-tier task means; derived by The Latent, not Roboflow’s official low-tier overall score.
Original source · Retrieved 2026-09-06
Matched exact metric/value and the archived benchmark-specific configuration review. Unknown native effort stays explicitly unknown.
Code Arena WebDev (overall): 1567 native-points
Contributes to the score. Source category: independent. Review status: reviewed-archive.
Configuration: gemini-3.8-flash-high
Metric: Published overall WebDev preference rating, not task success percentage or Arena Text score.
Original source · Retrieved 2026-09-06
Matched exact metric/value and the archived benchmark-specific configuration review. Unknown native effort stays explicitly unknown.
Cite this model record
The Latent. “Gemini 3.8 Flash: score and benchmark evidence.” 2026-09-06. Release tli-2026-v5.0-2026-09-06.
Download release data and definitions (JSON) · Data usage terms
Related reporting
These dated stories provide reporting context; they may describe an earlier release.