The LLM benchmark aggregator.

DeepSeek V4 Flash Vision

DeepSeek · 305B · open weights

Strongest in Data analysis (#9 of 57), weakest in Reasoning (#91 of 178). Above par in 9 of 13 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and GPT-6 Astra and ahead of Claude Opus 4.7 and Muse Spark 1.2.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Agents62−6.1#22/2682/5

Knowledge work65.7−7.9#19/1782/2

Core abilities55.9−11.5#37/2042/3

Data analysis60.2−5.9#9/571/1

Instruction following55.4−18.3#22/571/1

Language52.1−19.4#25/571/1

General intelligence55.8−14.3#56/2042/3

Coding47.8−21.3#89/1651/5

Agentic coding58.2−10.5#20/1571/4

Code generation31.2−38.8#76/771/2

Reasoning50−16.9#91/1781/4

Reasoning52.2−16.9#53/1401/3

Mathematics48−17.7#88/1401/2

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for DeepSeek V4 Flash Vision, left for the other.

Mistral Medium 3.5won 8 · lost 0

Mercury 2.5won 8 · lost 0

GPT-4.1 Nanowon 9 · lost 0

Mistral Large 3won 9 · lost 0

GPT-4.1 Miniwon 9 · lost 0

Llama 4 Maverick Instructwon 9 · lost 0

Grok Build 0.1won 13 · lost 0

Qwen3.6 27Bwon 12 · lost 1

Grok 4.3won 12 · lost 1

Gemini 3.5 Flash Litewon 12 · lost 1

GPT-5.4 Miniwon 12 · lost 1

Kimi K2.7 Codewon 12 · lost 1

MiniMax M3won 12 · lost 1

Inklingwon 12 · lost 1

GPT-4o Miniwon 9 · lost 1

DeepSeek V3won 9 · lost 1

Claude Haiku 4.5won 8 · lost 1

GPT OSS 120Bwon 8 · lost 1

GPT-4.1won 7 · lost 1

Mimo V2.5 Prowon 7 · lost 1

Qwen3.6 Pluswon 11 · lost 2

Kimi K2.6won 11 · lost 2

Ox Alphawon 9 · lost 2

Granite 4.2 3Bwon 7 · lost 2

Nemotron 3 Ultrawon 7 · lost 2

Gemini 2.5 Flashwon 7 · lost 2

Claude Sonnet 4.5won 7 · lost 2

GPT-5.4 Nanowon 10 · lost 3

Nemotron 3 Ultra 550B A55Bwon 10 · lost 3

Qwen3.8 27Bwon 10 · lost 3

Qwen3.6 35B A3Bwon 6 · lost 2

GLM-4.6won 6 · lost 2

K2 Horizon MoVA 36B A4Bwon 6 · lost 2

Kimi K2won 6 · lost 2

Gemma 4 31Bwon 6 · lost 2

GPT-5.2 Codexwon 8 · lost 3

GPT-6 Lunawon 9 · lost 4

GLM-5.3 Flashwon 9 · lost 4

GLM-5.2won 9 · lost 4

Muse Spark 1.1won 9 · lost 4

Qwen3.7 Maxwon 9 · lost 4

K2 Horizon 7Bwon 6 · lost 3

DeepSeek V3.2won 6 · lost 3

O3 Miniwon 5 · lost 3

K2 Horizon 0.9Bwon 5 · lost 3

K2 Horizon 375B A23Bwon 5 · lost 3

Gemini 3.1 Flash Litewon 5 · lost 3

GPT-5.6 Lunawon 8 · lost 5

Claude Sonnet 4.6won 8 · lost 5

DeepSeek V4 Flashwon 8 · lost 5

DeepSeek R1won 6 · lost 4

GPT-5.2won 7 · lost 5

Granite 4.2 8Bwon 5 · lost 4

K2 Horizon 3.7Bwon 5 · lost 4

O4 Miniwon 5 · lost 4

GPT-5 Miniwon 5 · lost 4

Kimi K2.5won 5 · lost 4

Union Alphawon 6 · lost 5

Claude Sonnet 5won 7 · lost 6

Gemini 3.6 Flashwon 7 · lost 6

Gemini 3.5 Flashwon 7 · lost 6

Muse Spark 1.2won 7 · lost 6

Claude Opus 4.7won 7 · lost 6

Gemma 4 12Bwon 4 · lost 4

Qwen3.5 9Bwon 4 · lost 4

Qwen3 235B A22B Instructwon 4 · lost 4

Claude Opus 4won 4 · lost 4

Claude Sonnet 4won 5 · lost 5

Claude Opus 4.5won 6 · lost 6

Claude Opus 4.8won 6 · lost 7

Qwen3.8 Flash Nextwon 6 · lost 7

DeepSeek V4 Prowon 6 · lost 7

Gemini 3.1 Prowon 6 · lost 7

GPT-5won 4 · lost 5

Gemini 2.5 Prowon 4 · lost 6

GPT-5.6 Terrawon 5 · lost 8

GPT-5.4won 5 · lost 8

Grok 4.5won 5 · lost 8

Claude Opus 4.6won 4 · lost 7

O3won 3 · lost 6

GLM-5.3won 4 · lost 9

Gemini 3.8 Flashwon 4 · lost 9

Gemini 3.7 Flashwon 4 · lost 9

DeepSeek V4.1 Flashwon 4 · lost 9

Grok 4.7won 4 · lost 9

GPT-5.5won 4 · lost 9

Grok 4won 2 · lost 6

Gemini 3 Prowon 2 · lost 6

Claude Opus 5won 3 · lost 10

Qwen3.8 Maxwon 3 · lost 10

Grok 4.6won 1 · lost 12

Kimi K3won 1 · lost 12

GPT-6 Solwon 1 · lost 12

Claude Opus 5.5won 1 · lost 12

Mimo V2.6 Prowon 0 · lost 8

GPT-5.6 Solwon 0 · lost 13

Muse Spark 1.3won 0 · lost 13

Claude Fable 5.1won 0 · lost 13

GPT-6 Astrawon 0 · lost 13

Claude Fable 5won 0 · lost 13

§ 3 · Sources

Where the numbers come from

4 publications, 11 figures. Every one links to the page it was read from.

Artificial Analysis ↗1 measureread 2026-09-26

Artificial Analysis Intelligence Index 35

GDPval-AA ↗1 measureread 2026-09-26

GDPval-AA 1534

AA-Briefcase ↗1 measureread 2026-09-26

AA-Briefcase 1428

LiveBench ↗8 measuresread 2026-09-26

LiveBench 76.8LiveBench · Reasoning 85.4LiveBench · Coding 68.2LiveBench · Agentic Coding 65.1LiveBench · Mathematics 87.8LiveBench · Data Analysis 79.5LiveBench · Language 80.4LiveBench · Instruction Following 71

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=deepseek-v4-flash-vision)](https://publicai.io/model-index/m/deepseek-v4-flash-vision)