An experimental vision-enabled DeepSeek V4 Flash model that adds image understanding while retaining the text, reasoning, coding, tool-calling, and agent capabilities of the base model. This route is served directly by DeepSeek, so privacy and logging guarantees are limited.
Added Aug 21, 2026
Context Window
1.0M
Max Output
384.0K
Input Price (Auto)
$0.44/1M
Output Price (Auto)
$1.32/1M
Cache Read (Auto)
$0.014/1M
Capabilities
Benchmarks
Benchmarks
Performance metrics and benchmarks
Sourced from LMArena.
Arena Score
1438.4
Overall Rank
#83 / 395
Votes
48,660
Confidence Interval
1434.2 - 1442.6
Category Scores
Coding
#96 / 390
14,304 votes
1479.4
Math
#73 / 379
2,458 votes
1441.2
Longer Query
#77 / 373
21,617 votes
1449.2
Creative Writing
#80 / 393
7,996 votes
1408.9
Instruction Following
#77 / 395
16,790 votes
1434.0
Hard Prompts
#86 / 395
32,470 votes
1458.3
Additional Categories22
Japanese
#56 / 261
598 votes
1422.7
Industry Legal And Government
#69 / 368
3,874 votes
1458.3
Chinese
#75 / 367
2,518 votes
1479.9
Industry Life And Physical And Social Science
#75 / 393
7,974 votes
1461.1
Expert
#77 / 345
4,909 votes
1467.5
Industry Medicine And Healthcare
#77 / 363
3,491 votes
1463.1
Industry Writing And Literature And Language
#77 / 394
11,837 votes
1421.4
French
#78 / 276
1,878 votes
1458.6
English
#80 / 395
21,740 votes
1449.9
Polish
#81 / 219
1,015 votes
1429.9
Spanish
#81 / 277
1,444 votes
1431.8
Exclude Ties
#82 / 395
36,997 votes
1435.7
German
#82 / 296
877 votes
1435.1
Industry Entertainment And Sports And Media
#82 / 393
10,632 votes
1406.1
Industry Mathematical
#82 / 374
2,601 votes
1443.3
Hard Prompts English
#83 / 393
15,131 votes
1465.1
Russian
#83 / 359
5,191 votes
1437.1
Multi Turn
#84 / 393
8,671 votes
1445.7
Industry Software And It Services
#85 / 395
20,036 votes
1473.2
Korean
#90 / 265
915 votes
1376.5
Non English
#90 / 395
26,918 votes
1422.7
Industry Business And Management And Financial Operations
#94 / 388
10,050 votes
1434.9
Published 2026-08-27 · Matched as deepseek-v4-flash-high-preview
LMArena DatasetProviders
Auto routing is available for this model. Explicit provider selection is not available.
Loading provider options…
Related text models
Compare DeepSeek V4 Flash Vision Exp with similar models from the same provider or model family.
DeepSeek V4 Flash Latest
deepseek/deepseek-v4-flash-latestCompatibility alias that routes to the newest dated DeepSeek V4 Flash release.
DeepSeek V4 Flash 0731 (Thinking)
deepseek/deepseek-v4-flash-0731:thinkingDeepSeek V4 Flash 0731 Thinking enables reasoning by default on the re-post-trained Mixture-of-Experts model with a 1M-token context window, built for coding, reasoning, and agent workflows.
DeepSeek V4 Flash 0731
deepseek/deepseek-v4-flash-0731DeepSeek V4 Flash 0731 is a re-post-trained Mixture-of-Experts model with a 1M-token context window, built for coding, reasoning, and agent workflows.
DeepSeek V4 Flash
deepseek/deepseek-v4-flashDeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with a 1M-token context window, built for fast inference, high-throughput workloads, reasoning, coding, and agent workflows.
DeepSeek V4 Flash (Thinking)
deepseek/deepseek-v4-flash:thinkingDeepSeek V4 Flash Thinking enables DeepSeek's reasoning mode on the efficiency-optimized Mixture-of-Experts model with a 1M-token context window, built for fast inference, high-throughput workloads, reasoning, coding, and agent workflows.
DeepSeek V4 Pro 0813
deepseek/deepseek-v4-pro-0813DeepSeek V4 Pro 0813 is the general-availability release of DeepSeek V4 Pro, built for coding, tool use, cybersecurity, automation, and long-horizon agent workflows. DeepSeek reports strong gains over the preview and scores above Opus 4.8 on Terminal Bench 2.1, Cybergym, DeepSWE, and AutomationBench. It supports a 1M-token context window.