DeepSeek V4 Flash Vision (exp)

deepseek/deepseek-v4-flash-vision-exp

fast hybrid-attention reasoning with vision.

cmd --model deepseek/deepseek-v4-flash-vision-exp
Intelligence index
40.7
Output speed
119.5 tok/s
Input
$0.15 /M
Output
$0.60 /M
Cache read
$0.003 /M
Agent-loop cost
$0.05 /M in
Context window
1M tokens
Released
August 21, 2026
Modalities
→

vs. the lineup

DeepSeek V4 Flash Vision (exp) beside its stablemates and nearest rivals. The ◆ marks the best value in each column across every row shown.

pin a rival:
ModelIntelligenceCodingSpeedInput $/MOutput $/MBlended $/MContext
Muse Spark 1.3 Contributor53◆76.3◆221◆$0.10◆$0.20◆$0.13◆1.05M◆
DeepSeek V4 Pro (latest)42.168.862.1$0.66$1.98$0.991M
DeepSeek V4 Flash (latest)4169.1117.7$0.15$0.60$0.261M
DeepSeek V4 Flash Vision (exp) ◆40.765119.5$0.15$0.60$0.261M
DeepSeek V4 Flash Fast———$0.28$0.56$0.351M
DeepSeek V4.1 Flash———$0.15$0.60$0.261M
Kimi K3pinned50.276.242.2$3$15$61M
Kimi K2.7 Codepinned32.760.865.7$0.95$4$1.71256K
Compare all 72 models →

coding performance

The Intelligence Index and its sub-scores, ranked against every scored model in the catalog. A metric that has not been measured for DeepSeek V4 Flash Vision (exp) has been left empty.

Coding Index
65
#33 of 53 scored
Terminal-Bench
74.2
#33 of 53 scored
Intelligence Index
40.7
#30 of 59 scored
Long-context reasoning
81.3
reasoning across a long context
SciCode
49.7
scientific coding
GPQA Diamond
91.3
graduate-level QA

peak hours

DeepSeek V4 Flash Vision (exp) is not billed at one flat rate — the price changes with the time of day. Every other price on this page is the off-peak rate, which is what a request costs 17 hours out of 24. Peak runs Monday to Friday only, so it is never charged at the weekend.

BandHours (UTC)Input $/MOutput $/MCache read $/M
Off-peak17h/day$0.15$0.60$0.003
Peak7h/day · 01–04 & 06–10 UTC, Mon–Fri$0.30$1.20$0.006

usage calculator

How far a month of credits goes on DeepSeek V4 Flash Vision (exp).

Input tokensfresh prompt
800
Output tokensmodel reply
180
Cache read tokensre-read context
50K
cost / request $0.0004 · in $0.15 · out $0.60 · cache $0.003 per M
fresh input 32%output 29%cache reads 40%
Requests / 30 days
26K
$10 credits ÷ $0.0004 per request
~5.3K quick fixes~1.1K bug fixes~176 feature PRs

what real work costs

Real coding tasks priced end to end on DeepSeek V4 Flash Vision (exp), from a quick lookup to a full-repo agent run.

One agent task
$0.01
180K in at 75% cache hit, 12K out
What you pay
$0.07 /M
all-in across every token that task touched
Sticker input
$0.15 /M
cache reads bill at $0.003 /M instead
TaskTokens in · outDeepSeek V4 Flash Vision (exp)Muse Spark 1.3 ContributorClaude Haiku 4.5
Quick lookup / one-liner8K · 1K$0.0007$0.0003$0.0065
Review a 500-line PR60K · 4K$0.0052$0.0027$0.04
Fix a bug (agent loop)180K · 12K$0.01$0.0072$0.12
Refactor a module320K · 20K$0.03$0.01$0.21
Full-repo agent run900K · 45K$0.06$0.03$0.49

frequently asked

How much does DeepSeek V4 Flash Vision (exp) cost?
$0.15/M input and $0.60/M output, cache reads $0.003/M. In an agent loop most input is cache-read, so the effective input rate is about $0.05/M.
Which plan do I need?
Available on Go and above.
How do I switch to it?
Run cmd --model deepseek/deepseek-v4-flash-vision-exp, or type /model in a session and pick it. You can switch mid-session without losing context.
DeepSeek V4 Pro (latest)→
hybrid-attention long-context reasoning · $0.66/$1.98 per M
DeepSeek V4 Flash (latest)→
fast hybrid-attention reasoning · $0.15/$0.60 per M
All supported models→
sortable by score, speed and price

Ship code that matches your taste

Command Code is the AI coding agent that continuously learns your taste. Start for $1.

Start nowSee pricing
Benchmarks from Artificial Analysis (v4.3)commandcode.ai/models/deepseek-v4-flash-vision-exp