VerdictPal
Join

Models / DeepSeek

DeepSeek V4 Pro (Reasoning, Max Effort)

DeepSeek V4 · Released 2026-04-24

AA Intelligence Index 44.3. Listed on Artificial Analysis with API pricing $0.435/$0.87 per 1M tokens. Benchmark rows sync from the AA snapshot — see the ledger on this page.

Frontier model tracked on Artificial Analysis. Use the benchmark ledger below for sourced scores; we do not infer numbers beyond published rows.

#17 of 51 on AA Index · current snapshot
44 AA Index · 10 benchmark rows · 1 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite57
Context128K
Input / 1M$0.43
Output / 1M$0.87
Knowledge cutoffopen
Output speed73 t/s
TTFT (default API)1.01s
Statussolid
ReasoningCodingAgenticLong context
Reasoning
89
Coding
50
Agentic
64
Long context
66

67average across 4 tested families

Benchmark placements

Where DeepSeek V4 Pro (Reasoning, Max Effort) places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
IFBenchReasoning#10/ 5776Artificial Analysis2026-07-21
Artificial Analysis Intelligence IndexReasoning#22/ 6744Artificial Analysis2026-07-21
Humanity's Last ExamReasoning#25/ 6536Artificial Analysis2026-07-21
GPQA DiamondReasoning#28/ 6989Artificial Analysis2026-07-21
SciCodeCoding#28/ 6550Artificial Analysis2026-07-21
Terminal-BenchAgentic#32/ 6964Artificial Analysis2026-07-21
AA-LCRLong context#33/ 6466Artificial Analysis2026-07-21
τ³-Bench BankingAgentic#33/ 6426Artificial Analysis2026-07-21
AA output speedPerformance#38/ 69Artificial Analysis2026-07-21
AA time to first tokenPerformance#45/ 69Artificial Analysis2026-07-21
Family context

The three highest-scoring carded models in each capability family. Where DeepSeek V4 Pro (Reasoning, Max Effort) shows up, it's highlighted.

ReasoningDeepSeek V4 Pro (Reasoning, Max Effort) · 89
  1. 1
    Claude Mythos Preview
    95
  2. 2
    GPT-5.6 Sol
    94
  3. 3
    Kimi K3
    94
CodingDeepSeek V4 Pro (Reasoning, Max Effort) · 50
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticDeepSeek V4 Pro (Reasoning, Max Effort) · 64
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextDeepSeek V4 Pro (Reasoning, Max Effort) · 66
  1. 1
    Kimi K3
    75
  2. 2
    GPT-5.1
    75
  3. 3
    GPT-5.6 Sol
    74
Newest receipts

Benchmark rows added to the public ledger for DeepSeek V4 Pro (Reasoning, Max Effort) in the last 120 days. Older rows live in the full table below.

10 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context66.33%Artificial Analysis2
AA output speedPerformance73 t/sArtificial Analysis2
AA time to first tokenPerformance1.01sArtificial Analysis2
Artificial Analysis Intelligence IndexReasoning44.3%Artificial Analysis2
GPQA DiamondReasoning88.8%Artificial Analysis2
Humanity's Last ExamReasoning35.9%Artificial Analysis2
IFBenchReasoning76.46%Artificial Analysis2
SciCodeCoding50%Artificial Analysis2
τ³-Bench BankingAgentic25.77%Artificial Analysis2
Terminal-BenchAgentic64.04%Artificial Analysis2
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Reasoning89
Coding50
Agentic64
Long context66
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover or tab any dot for name, score, and input price.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMiniMaxxAIMetaZhipuXiaomiMeituanCohere

Full benchmark ledger

Every catalog benchmark for DeepSeek V4 Pro (Reasoning, Max Effort). Scores link to the original source; gaps mean no public row exists yet.

10 sourced rows

11 of 32 catalog benchmarks have a sourced row for DeepSeek V4 Pro (Reasoning, Max Effort).

Reasoning

Coding

Agentic

Long context

Performance

Sources