LuminaBench
Agentic benchmark / Arena

Agent Arena

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.

Observed agent orchestrator effects in real user workflows.

Version Aggregate · 12 Oct 2026 signalsVerified 2 November 202657 published resultsOfficial source

Agent Arena · Lumina Score

Aggregate IPS · Published 2026-10-28

LuminaBench
Show
20 of 55 models · best variant per model
03006009001,2001,180Claude Opus 5.5Max1,138Gemini 4 CarbonHigh1,119Claude Sonnet 5.5Max1,115Claude Fable 5.1Max1,112GPT-6 AstraMax1,108GPT-6.1 SolMax1,093Gemini 4 ArgonHigh1,091GPT-6.1 TerraMax1,087Muse Wonder FullMax1,070Claude Fable 5High1,069GPT-6 SolMax1,066Claude Opus 5Max1,052Kimi K3.1 ProMax1,042GPT-5.6 SolxHigh1,028Muse WonderMax1,026Claude Sonnet 5High1,025MiMo-V2.6-Pro1,024DeepSeek-V4.1-Flash-20261017Max1,024Gemini 3.8 FlashHigh1,023Muse Spark 1.3Max

Signed publisher IPS estimates and 95% bounds; observations and sessions retained. Missing metrics are omitted from their graph; quality scores remain available.

Results table

Source & conversion

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

Signed IPS causal estimates with publisher 95% intervals, observation and session counts. No cost, time, token or tool-call measurements are published.

How the Lumina conversion works

1,000 + 75 × ((original score − -0.0010993491128136034) / 0.08770085980286019). 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
agent-arena-z75-v1 · experimental
Reference mean
-0.0010993491128136034
Population SD
0.08770085980286019
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.

0 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 · Agent methodology · Current signals

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