LuminaBench
Google / Tested model

Gemini 3.8 Flash

Benchmark results and original creations.

3 tested configurations9 test results · 3 categoriesReleased 2 September 2026
Best recorded results

Benchmark scores

Best eligible score in each benchmark.

API pricing

USD / 1M tokens
Input
$0.75
Output
$3.75
Cached input
$0.075
API-equivalent rates in Cursor · Checked 30 Oct 2026Source
Pricing details

Base introductory API rates shown by Cursor. The existing provider snapshot lists an increase from 1 January 2027.

Gemini 3.8 Flash High: Voxel Pagoda front view.

Voxel

By LuminaDesigned and run by LuminaBench. Original outputs and evidence are retained.

3 configurations

866Best score · High · Antigravity
Explore Voxel
Gemini 3.8 Flash · High: Seaplane front view

3D

By LuminaDesigned and run by LuminaBench. Original outputs and evidence are retained.

3 configurations

945Best score · Low · Antigravity
Explore 3D
Gemini 3.8 Flash high: character selection

GameDev

By LuminaDesigned and run by LuminaBench. Original outputs and evidence are retained.

3 configurations

949Best score · High · Antigravity
Explore GameDev
CursorBench

CursorBench

Version 4.0
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.

2 configurations · Publisher-run · Experimental conversion

1,006Best score · High
Explore CursorBench
Terminal-Bench

Terminal-Bench

Version 4.0
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.

1 configuration · Publisher-run · Experimental conversion

906Best score · High
Explore Terminal-Bench

Coding Index

Version 2
By LuminaIndex designed by LuminaBench using attributed, publisher-run benchmarks. Lumina did not rerun those evaluations.

2 configurations · Derived from published benchmarks · Experimental · 2/2 Complete

956Best score · High
Explore Coding Index