Back

DeepSeek

DeepSeek V3.2

ReasoningOpen weightsModels.devHugging Face
Intelligence
16.0
Rank 296 of 744 measured models
Coding
44.2
Rank 141 of 298 measured models
Price/1M
$0.315
Rank 213 of 659 measured models
Context window
128K
Rank 773 of 1199 measured models

DeepSeek V3.2 (Non-reasoning) scores 16.0 on the Artificial Analysis Intelligence Index, ranking 296th of 744 measured models. The closest models on this index are Gemini 2.5 Pro (16.1), MiMo-V2-Flash (16.0), and Gemini 3.1 Flash-Lite (15.6). At $0.315 per million tokens (3:1 input/output blended price), it is cheaper than Gemini 2.5 Pro ($3.44) and Gemini 3.1 Flash-Lite ($0.563).

Artificial Analysis Indices
Non-reasoning
Intelligence16.0
Rank 296 of 744 measured models
Coding44.2
Rank 141 of 298 measured models
Math59.0
Rank 119 of 270 measured models
Reasoning levels
Each point is a reasoning level of this model: more reasoning often raises the score but costs more. Select a point to open that level.
Intelligence Indexmost efficient ↗Price / 1M tokens
Standard Benchmarks
MMLU Pro83.7%
Rank 47 of 349 measured models
GPQA Diamond75.1%
Rank 269 of 619 measured models
HLE11.2%
Rank 307 of 664 measured models
LiveCodeBench59.3%
Rank 118 of 343 measured models
AIME 202559.0%
Rank 119 of 270 measured models
IFBench49.0%
Rank 189 of 450 measured models
AA-LCR45.7%
Rank 342 of 575 measured models
TerminalBench Hard32.6%
Rank 100 of 432 measured models
Additional benchmarks8
Design Arena · 3d49.4%
Design Arena · Asciiart40.5%
Design Arena · Codecategories49.3%
Design Arena · Dataviz48.2%
Design Arena · Gamedev46.5%
Design Arena · Svg40.8%
Design Arena · Uicomponent46.8%
Design Arena · Website50.2%
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy