VerdictPal
Join

Models / DeepSeek

DeepSeek V4 Flash 0731 (Reasoning, Max Effort)

DeepSeek V4 · Released 2026-07-31

AA Intelligence Index 51.8. Listed on Artificial Analysis with API pricing $0.44/$1.32 per 1M tokens after the 17 Aug 2026 list-price change. 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.

#22 of 168 on AA Index · current snapshot
52 AA Index · 9 benchmark rows · 1 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite60
Context128K
Input / 1M$0.44
Output / 1M$1.3
Knowledge cutoffopen
Output speed106 t/s
TTFT (default API)0.96s
Statussolid

Family profile

Best published score in each covered benchmark family.

74/100 avg
Reasoning
91
Coding
50
Agentic
79
Long context
74

4 tested benchmark families

Benchmark placements

Where DeepSeek V4 Flash 0731 (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
τ³-Bench BankingAgentic#19/ 10639Artificial Analysis2026-09-01
Terminal-BenchAgentic#28/ 18479Artificial Analysis2026-09-01
Artificial Analysis Intelligence IndexReasoning#31/ 18752Artificial Analysis2026-09-01
GPQA DiamondReasoning#34/ 18591Artificial Analysis2026-09-01
AA output speedPerformance#37/ 187Artificial Analysis2026-09-01
Humanity's Last ExamReasoning#40/ 18039Artificial Analysis2026-09-01
SciCodeCoding#45/ 18050Artificial Analysis2026-09-01
AA-LCRLong context#55/ 17974Artificial Analysis2026-09-01
AA time to first tokenPerformance#66/ 186Artificial Analysis2026-09-01
Family context

The three highest-scoring models with pages in each capability family. Where DeepSeek V4 Flash 0731 (Reasoning, Max Effort) shows up, it's highlighted.

ReasoningDeepSeek V4 Flash 0731 (Reasoning, Max Effort) · 91
  1. 1
    Claude Mythos Preview
    95
  2. 2
    Gemini 3.7 Flash
    95
  3. 3
    Grok 4.6
    95
CodingDeepSeek V4 Flash 0731 (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 Flash 0731 (Reasoning, Max Effort) · 79
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextDeepSeek V4 Flash 0731 (Reasoning, Max Effort) · 74
  1. 1
    Muse Spark 1.2
    83
  2. 2
    Kimi K3
    83
  3. 3
    Gemini 3.5 Flash
    81
Newest receipts

Benchmark rows added to the public ledger for DeepSeek V4 Flash 0731 (Reasoning, Max Effort) in the last 120 days. Older rows live in the full table below.

9 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context74.33%Artificial Analysis5
AA output speedPerformance106 t/sArtificial Analysis5
AA time to first tokenPerformance0.96sArtificial Analysis5
Artificial Analysis Intelligence IndexReasoning51.8%Artificial Analysis5
GPQA DiamondReasoning90.8%Artificial Analysis5
Humanity's Last ExamReasoning38.6%Artificial Analysis5
SciCodeCoding49.9%Artificial Analysis5
τ³-Bench BankingAgentic39.38%Artificial Analysis5
Terminal-BenchAgentic78.65%Artificial Analysis5
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.

Reasoning91
Coding50
Agentic79
Long context74
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMetaMiniMaxxAIZhipuXiaomiOtherMeituanCohereNVIDIA

Full benchmark ledger

Every catalog benchmark for DeepSeek V4 Flash 0731 (Reasoning, Max Effort). Scores link to the original source; gaps mean no public row exists yet.

9 sourced rows

9 of 37 catalog benchmarks have a sourced row for DeepSeek V4 Flash 0731 (Reasoning, Max Effort).

Reasoning

Coding

Agentic

Long context

Performance

Sources