Inference
EN
DeepSeek

DeepSeek V4 Pro 0423

DeepSeek V4 Pro 0423 is a model from DeepSeek. 24 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Input: text
  • Released 2026-04
RankedEnough results for an index.
IndexAll categories, out of 100.57.6±4.8Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.89/sMedian 67/s
First tokenSeconds to the first output token.0.7 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.892Median $0.60
Output / 1MUS dollars per 1M output tokens.$1.78Median $2.55
ContextMaximum tokens in one request.1.05MMedian 262K
EloLMArena rating and rank.1451±4 #50Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
deepseek/deepseek-v4-pro
Max outputMaximum output tokens per request.
384K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.074
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
8 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
OpenRouterRouter$0.955$1.911.05M
Deep InfraHost$1.30$2.601.05M
GMI CloudHost$1.39$2.781.05M
SiliconFlowHost$1.50$3.141M
Novita AIHost$1.60$3.201.05M
Arcee AIDeprecatedHost$1.74$3.48512K
DigitalOceanCloud$1.74$3.481.05M
Microsoft AzureCloud$1.74$3.481M

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Run locally

1598.84B parameters
16-bit
3358.6 GB
8-bit
1784.7 GB
4-bit
1018.8 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy needs more than 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.4 counted55.997 / 218
CodingCode writing and repair.7 counted59.274 / 209
ReasoningLogic problems and puzzles.3 counted54.686 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.4 counted59.965 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted52.7—
MathMath problems.5 counted57.653 / 168

Results

24 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
LiveCodeBenchCoding87.5% ±1.013 / 142
SWE-Bench verified · Epoch AICoding77.6% ±1.95 / 32
OTIS Mock AIME 2024-2025Math96.7% ±2.036 / 200
GPQA diamond · Epoch AIKnowledge89.6% ±1.744 / 217
Terminal-Bench 2.0 · Vals AIAgentic56.2% ±5.314 / 66
GPQA Diamond · Vals AIKnowledge89.4% ±1.730 / 135
MMLU ProKnowledge87.2% ±0.332 / 135
IOI v1Coding35.8% ±7.115 / 61
SWE-benchCoding77.4% ±1.935 / 87
SimpleQA VerifiedKnowledge47.0% ±1.637 / 91
LiveBench MathematicsMath90.7%31 / 66
LiveBench Data AnalysisReasoning74.5%41 / 80
Vibe Code Bench v1.1Coding49.9% ±4.866 / 123
LiveBench LanguageKnowledge78.1%37 / 66
Code MigrationCoding26.2% ±4.046 / 77
SkillsBenchCoding51.3% ±4.621 / 35
LiveBench ReasoningReasoning82.7%51 / 80
Mystery Game PuzzlesReasoning17.0% ±3.851 / 80
FrontierMath-Tiers-1-3-v2-PrivateMath45.3% ±3.058 / 90
Terminal-Bench 4.0Agentic11.1% ±1.344 / 61
Chess PuzzlesReasoning19.0% ±3.968 / 90
Terminal-Bench 2.1 · Vals AIAgentic50.2% ±1.558 / 75
LiveBench Instruction FollowingInstruction62.4%52 / 66
ProofBench v1.1Math16.0% ±3.746 / 54
LiveBench Agentic CodingAgentic42.6%57 / 66
LiveBench CodingCoding70.0%59 / 66
FrontierMath-Tier-4-v2-PrivateMath2.4% ±2.465 / 72
Finance Agent v1.1Finance60.4% ±2.84 / 51
CorpFin v2Finance61.4% ±1.052 / 131
TaxEval v2Finance72.1% ±0.968 / 143
CaseLaw v2Legal59.4% ±1.527 / 53
Public Benefits Bench v1.1Public benefits62.9% ±1.326 / 50
ProgramBenchCoding0.0%34 / 65
Tax Agent BenchFinance25.2% ±3.144 / 82
Public Benefits Bench v1Public benefits57.6% ±1.37 / 13
LegalBenchLegal80.3% ±0.587 / 149
MedCodeHealthcare40.5% ±2.164 / 106
EMBFinance51.6% ±3.155 / 75
Legal Research BenchLegal23.1% ±2.966 / 90
Finance Agent (v2)Finance44.1% ±0.758 / 79
MedScribeHealthcare75.1% ±2.081 / 109

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYLiveBenchNo licence statedVals AINo licence statedmodels.devMIT

History

5 releases
Index58.6 → 57.6
23 Oct 202612 Nov 2026
Output / 1M$0.418 → $1.78
23 Oct 202612 Nov 2026

Questions

How much does DeepSeek V4 Pro 0423 cost?
$0.892 per 1M input tokens, $1.78 per 1M output tokens.
How good is DeepSeek V4 Pro 0423?
Its index is 57.6 out of 100. That is place 88 of 283 ranked models.
What is DeepSeek V4 Pro 0423 best at?
Knowledge: 59.9 out of 100, place 65 of 268.
How fast is DeepSeek V4 Pro 0423?
89 output tokens per second, 0.7 s to the first token.
What is the context window of DeepSeek V4 Pro 0423?
1.05M tokens.
Is DeepSeek V4 Pro 0423 open source?
Yes. Its weights are published.
When was DeepSeek V4 Pro 0423 released?
April 2026.

Badge

DeepSeek V4 Pro 0423 on Inference360: place 88 of 283, index 57.6
HTML
<a href="https://inference360.ai/models/deepseek-deepseek-v4-pro-0423"><img src="https://inference360.ai/badges/deepseek-deepseek-v4-pro-0423.svg" alt="DeepSeek V4 Pro 0423 on Inference360: place 88 of 283, index 57.6" height="20"></a>
Markdown
[![DeepSeek V4 Pro 0423 on Inference360: place 88 of 283, index 57.6](https://inference360.ai/badges/deepseek-deepseek-v4-pro-0423.svg)](https://inference360.ai/models/deepseek-deepseek-v4-pro-0423)

Compare with

Gemini 3 Flash Preview57.5Qwen3.7 Plus57.7Kimi K2.7 Code57.7DeepSeek V4 Flash 073161.9DeepSeek V3.244.2

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeGLM-5.3-Flash63.1 · $0.078/1M

More from DeepSeek

All models
DeepSeek V4 Flash 0423UnrankedDeepSeek V4 Flash 073161.9DeepSeek V4 Flash LatestUnrankedDeepSeek V4 Pro 081365.5DeepSeek V4 Flash Vision Exp60.4