DeepSeek V4 Pro
DeepSeek · 3 thinking configurations
Intelligence30.4
Input / 1M$0.44
Output / 1M$0.87
Context1M
Tariff: DeepSeek · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost1 measured configurations · 2 without both values
Selected: thinking max · Click a point to switch configuration
Open MoE flagship with million-token context for coding and long agent runs
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
30.4 pointsDeepSWE · pass@1Software-engineering tasks on mini-swe-agent v1.1. Pass@1 across evaluated runs, with measured cost per task. The 95% intervals may overlap.
62.8 % ±6.3GPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
88.8 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
37.5 %SciCodeCode generation for scientific research problems.
50.8 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
64.0 %τ²-BenchTool use and interaction in assistance scenarios. Does not replace an evaluation of your integration.
96.2 %IFBenchVerifiable instruction following. A proxy for precision rather than a benchmark of creative style.
76.5 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
43.0 %Specifications from the catalog
Inputtext
Outputtext
Tool callingYes
Structured outputYes
Speed84.2 tok/s
First token1.52 s
Maximum output16.4K
Catalog listing date2026-04-24
DeepSWE v1.1 · 113 tasks · mini-swe-agent · cost $0.24 / task · 155 mean steps · 106K mean output tokens. Published by the source 9/22/2026.