LuminaBench
Coding benchmark / SWE-bench authors

SWE-bench Multilingual — Original

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.

Resolved

Version SWE-bench-Multilingual-193160a463a4Verified 6 November 202613 published resultsOfficial source

SWE-bench Multilingual — Original

LuminaBench
Show
13 of 13 models · best variant per model
03006009001,2001,065Gemini 3 Flash1,059Claude 4.6 Opus1,047Claude 4.5 Opus1,038GLM-51,028Gemini 3 Pro1,025MiniMax M2.51,016Kimi K2.51,013Claude 4.5 Sonnet1,010GPT-5.2High1,006GPT-5.2-Codex992Claude 4.5 Haiku940DeepSeek V3.2763GPT-5 mini

Lumina Score vs Cost per Task

No comparable per-task resource measurements were published.

LuminaBench
13 of 13 families
Lower resource use and higher score (green); higher resource use and lower score (red). Guides split at the visible selection’s medians.$0.1$0.3$17008009001,0001,100Cost per task · USD (log scale)Lumina ScoreGemini 3 FlashClaude 4.6 OpusGPT-5 miniClaude 4.5 OpusGLM-5MiniMax M2.5

No comparable per-task resource measurements were published. Missing metrics are omitted from their graph; quality scores remain available.

Results table

Method & coverage

Original Multilingual submission leaderboard; exact complete model+agent systems, checked flags and warnings preserved. No multimodal results imported; submissions are not interchangeable native-model measurements.

No comparable per-task resource measurements were published.

How the Lumina conversion works

1,000 + 75 × ((original score − 65.6) / 8.193900170248599). The result is rounded once. Scores may tie.

Frozen initial observed text-only cohort; one mean per exact publisher release/API, balancing repeated settings/routes within that release. Original score and protocol remain separate. No inferred variants or future refitting.

Calibration
swe-bench-multilingual-original-z75-v1 · experimental
Reference mean
65.6
Population SD
8.193900170248599
Coverage
Standalone benchmark.
Identity, exclusions & attribution

Original measurements and exact source settings retained; unknown API, effort and harness versions remain unknown.

Text software issue descriptions and code/terminal agent tools; Multimodal board excluded.

Standalone text benchmark; no new combined index. Source identity/effort/protocol qualifications remain in Data and CSV.

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 source

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