Back

DeepSeek

DeepSeek V4 Pro 0424

ReasoningOpen weights
Intelligence
30.4
Rank 122 of 744 measured models
Coding
59.4
Rank 87 of 298 measured models
Output speed
83.4 t/s
Rank 217 of 403 measured models
Price/1M
$0.544
Rank 294 of 659 measured models
Context window
1M
Rank 84 of 1199 measured models
Cost per task
$0.122
Rank 58 of 193 measured models

DeepSeek V4 Pro 0424 (Max) scores 30.4 on the Artificial Analysis Intelligence Index, ranking 122nd of 744 measured models. The closest models on this index are K2 Horizon 375B A23B (30.5), GPT-5.2 (30.4), and Claude Sonnet 4.6 (batch) (30.1). At $0.544 per million tokens (3:1 input/output blended price), it is cheaper than GPT-5.2 ($4.81). With an output speed of 83.4 t/s, it is faster than GPT-5.2 (73.3 t/s), slower than K2 Horizon 375B A23B (474 t/s).

Artificial Analysis Indices
Maximum
Intelligence30.4
Rank 122 of 744 measured models
Coding59.4
Rank 87 of 298 measured models
Agentic26.3
Rank 78 of 189 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
Capability indexes
Performance in specific industries, measured by Artificial Analysis
Finance & Accounting34.5
Rank 99 of 218 measured models
Strategy & Ops39.3
Rank 90 of 218 measured models
Legal33.7
Rank 106 of 218 measured models
Healthcare & Medical33.0
Rank 48 of 82 measured models
Engineering31.6
Rank 106 of 217 measured models
Economics37.3
Rank 119 of 225 measured models
Standard Benchmarks
GPQA Diamond88.8%
Rank 88 of 619 measured models
HLE37.5%
Rank 102 of 664 measured models
SciCode50.8%
Rank 100 of 231 measured models
IFBench76.5%
Rank 21 of 450 measured models
AA-LCR74.7%
Rank 160 of 575 measured models
TerminalBench Hard46.2%
Rank 30 of 432 measured models
Terminal-Bench 2.164.0%
Rank 91 of 244 measured models
Terminal-Bench v414.6%
Rank 69 of 218 measured models
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