LuminaBench
Coding benchmark / Arena

Arena WebDev

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.

Published preferences for generated web applications.

Version OverallVerified 2 November 2026150 published resultsOfficial source

Arena WebDev · Lumina Score

Overall preferences · Published 2026-10-28

LuminaBench
Show
20 of 138 models · best variant per model
03006009001,2001,171Claude Opus 5.5Max1,153GPT-6 AstraMax1,151Claude Sonnet 5.5xHigh1,150Gemini 4 CarbonHigh1,142GPT-6.1 SolMax1,139Claude Fable 5.1Max1,129Muse Wonder FullMax1,114Kimi K3.1 ProMax1,112Claude Opus 5Max1,106Gemini 4 ArgonHigh1,106GPT-6 SolMax1,105Qwen3.8-Max (0902)1,103GPT-6.1 TerraMax1,103Qwen3.8-Max1,099DeepSeek-V4.1-Flash-20261017Max1,095Kimi K3Max1,095Muse Spark 1.3Max1,093Muse WonderMax1,086Grok 4.7xHigh1,086MiMo-V2.6-Flash

Publisher preference ratings, bounds and vote counts retained. Missing metrics are omitted from their graph; quality scores remain available.

Results table

Source & conversion

Arena webdev, overall only. Published 2026-10-28; underlying evaluation timestamps are not supplied.

Preference ratings with exact publisher bounds, variance and vote counts. The dataset does not specify the interval level. No resource telemetry is published.

How the Lumina conversion works

1,000 + 75 × ((original score − 1463.8447770056782) / 151.92111567407474). The result is rounded once. Scores may tie.

Frozen initial valid overall cohort; one arithmetic mean per exact evaluated release/system, equal weight across releases rather than effort rows. Unbounded rating/IPS scales use a broad finite numeric guard, not percentage bounds or sign clamping. Subsequent refreshes never refit this calibration. Conversion is not an estimate or combined index.

Calibration
arena-webdev-z75-v1 · experimental
Reference mean
1463.8447770056782
Population SD
151.92111567407474
Coverage
Standalone benchmark.
Identity, exclusions & attribution

Exact publisher labels and release qualifiers are retained. Only explicit native effort tokens create reasoning variants; token budgets remain separate.

Agent orchestrator, routed, dated and harness-qualified systems are not automatically equated with naked base models.

2 ambiguous repeated records are retained in original.json but excluded from ranking.

All published reasoning settings, including Claude Max, are included by default. Display labels use Low and xHigh; source labels remain in the retained records.

Official dataset · Pinned source revision · CC BY 4.0

Publisher terms. Scores and resource figures are attributed to the original publisher.