Skip to content

Home Leaderboard Compare Models Benchmarks News

Search LUMINA BENCH

Navigate the site or open source-checked model and benchmark registry records.

Search models, benchmarks and pages

ESC

Navigate

Source-qualified snapshot · Power Index 1.6.0↑↓ navigate · Enter open

Latest rankings

Navigation

Product

Home→Leaderboard→Compare→Models→Benchmarks→News→

Research & company

Methodology Data sources Changelog About

Latest rankings

Independent rankings. Commercial relationships never affect scoring.

Loading table

Loading LUMINA BENCH

Research briefing

Signals worth opening.

A future briefing for verified ranking changes and methodology updates. Preview only—nothing is transmitted or stored.

Email address

Independence standard

Commercial relationships never affect rankings, verification, or editorial conclusions.

Read sponsor disclosure

Independent power rankings from hard benchmarks — speed and price stay decision context, not rank inputs.

Product

Leaderboard
Compare
Models
Benchmarks

Research

How rankings work
Data sources
Changelog

Company

About
News
Contact
Sponsor disclosure

Legal

Privacy
Terms
Advertise

© 2026 LUMINA BENCH. Evidence before rank.Data refreshed 21 Jul 2026 · Power Index 1.6.0 · 820 models

Home
Models
DeepSeek V3.2 (Reasoning)

DeepSeek · stable

DeepSeek V3.2 (Reasoning)

DeepSeek V3.2 (Reasoning) is a model variant tracked by the Artificial Analysis Free API under DeepSeek.

Not ranked · Not evaluated score · Confidence Not graded

Loading DeepSeek V3.2 (Reasoning) profile