// head_to_head

DeepSeek V4 Flash Vision Exp vs DeepSeek V4 Pro 0813

Published LiveBench scores across all seven categories, live list pricing, context windows, and the measured cost of a point of capability — for both models, side by side.

OpenRouter + LiveBenchAll comparisonsFull leaderboard

Effectively the same quality — DeepSeek V4 Flash Vision Exp is the cheaper way to get it.

The two are within 0.7 points overall, which is inside the range that effort settings alone move a LiveBench score, so treat them as quality-equivalent. DeepSeek V4 Flash Vision Exp lists 5.4× cheaper per blended million tokens. When quality ties, cost is the whole decision.

deepseek

DeepSeek V4 Flash Vision Exp

Blended / 1M
$0.330
Context
1.0M
Released
Aug 21, 2026
Overall score
76.8
reasoningtool callingimage inputprompt caching

deepseek

DeepSeek V4 Pro 0813

Blended / 1M
$1.78
Context
1.0M
Released
Aug 12, 2026
Overall score
77.4
reasoningtool callingprompt caching

Specs and pricing

MetricDeepSeek V4 Flash Vision ExpDeepSeek V4 Pro 0813
LiveBench overall

Mean of the seven LiveBench category scores, 0–100. Higher is better.

76.877.4
Cost per point

Measured benchmark spend divided by overall score — dollars per point of capability.

$0.0277$0.0241
Blended price / 1M

3:1 input:output mix, the usual shape of production traffic.

$0.330win$1.78
Input price / 1M$0.220win$1.19
Output price / 1M$0.660win$3.56
Cached input / 1M

Price of an input token served from the prompt cache, where the provider publishes one.

$0.0070win$0.040
Context window1.0M1.0M
Max output tokens384K

Benchmarks by category

An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — DeepSeek V4 Flash Vision Exp on top, DeepSeek V4 Pro 0813 below, both out of 100.

Agentic coding
65.1
54.9
Coding
68.2
77.2
Reasoningtoo close to call
85.4
85.8
Mathematics
87.8
95.1
Data analysistoo close to call
79.5
79.2
Language
80.4
82.1
Instruction following
71.0
67.7

What each one costs to run

Per-token prices are hard to feel. These are monthly list costs for both models across five workload shapes, using each provider's published cached-input rate where there is one.

WorkloadDeepSeek V4 Flash Vision ExpDeepSeek V4 Pro 0813
Support chatbot

1.2K in / 400 out × 200K requests

$90.26/mo$487.56/mo
RAG assistant

8K in / 600 out × 100K requests

$130.40/mo$704.88/mo
Coding agent

40K in / 4K out × 20K requests

$109.52/mo$592.42/mo
Document extraction

20K in / 1.5K out × 50K requests

$258.85/mo$1397.88/mo
Bulk classification

500 in / 20 out × 5M requests

$509.50/mo$2752.20/mo
Run these two through the cost calculator

Which should you pick?

The workload is coding or agentic work

DeepSeek V4 Flash Vision Exp

Leads on agentic coding — 65.1 against 54.9.

DeepSeek V4 Flash Vision Exp vs DeepSeek V4 Pro 0813 FAQ

Which is better, DeepSeek V4 Flash Vision Exp or DeepSeek V4 Pro 0813?

Effectively the same quality — DeepSeek V4 Flash Vision Exp is the cheaper way to get it. The two are within 0.7 points overall, which is inside the range that effort settings alone move a LiveBench score, so treat them as quality-equivalent. DeepSeek V4 Flash Vision Exp lists 5.4× cheaper per blended million tokens. When quality ties, cost is the whole decision.

Is DeepSeek V4 Flash Vision Exp cheaper than DeepSeek V4 Pro 0813?

DeepSeek V4 Flash Vision Exp is cheaper. On a 3:1 input:output blend, DeepSeek V4 Flash Vision Exp lists at $0.330 per million tokens and DeepSeek V4 Pro 0813 at $1.78 — DeepSeek V4 Flash Vision Exp is 5.4× cheaper. Input and output are priced separately — DeepSeek V4 Flash Vision Exp charges $0.220 in and $0.660 out, DeepSeek V4 Pro 0813 charges $1.19 and $3.56 — so the model that looks cheaper flips depending on how output-heavy your workload is.

DeepSeek V4 Flash Vision Exp vs DeepSeek V4 Pro 0813: which scores higher on benchmarks?

DeepSeek V4 Flash Vision Exp scores 76.8 and DeepSeek V4 Pro 0813 scores 77.4 overall on LiveBench, the mean of its seven categories. That gap is inside the range that effort settings alone move a score, so treat them as equivalent on published quality. Category scores differ from the overall figure — a model can lead on reasoning and trail on coding, which the per-category table above breaks out.

Which gives better value for money, DeepSeek V4 Flash Vision Exp or DeepSeek V4 Pro 0813?

They are close. DeepSeek V4 Flash Vision Exp costs $0.0277 per point of overall capability and DeepSeek V4 Pro 0813 costs $0.0241, a difference small enough that workload shape will matter more than the rate.

Does DeepSeek V4 Flash Vision Exp or DeepSeek V4 Pro 0813 have a bigger context window?

They are effectively the same — 1.0M for DeepSeek V4 Flash Vision Exp and 1.0M for DeepSeek V4 Pro 0813.

Do DeepSeek V4 Flash Vision Exp and DeepSeek V4 Pro 0813 support prompt caching?

Both publish a cached-input rate: $0.0070 per million for DeepSeek V4 Flash Vision Exp and $0.040 for DeepSeek V4 Pro 0813, against full input rates of $0.220 and $1.19. On a workload with a long stable prefix — a system prompt, a tool schema, a retrieved corpus — that changes the economics more than the headline price does.

Related comparisons

How these numbers are produced

  • Price — provider list price from OpenRouter, refreshed every 15 minutes. “Blended” is a 3:1 input:output mix.
  • ScoresLiveBench release 2026-06-25, using their own category map. Each model shows its strongest published run. A blank means “not evaluated”, never “bad”.
  • Cost per point — the measured dollars LiveBench spent on the run, divided by the score it earned.
  • “Win” — awarded only past a threshold: one full point on a benchmark score, 10% on a price, 25% on a context window. Anything tighter reports as a tie, because effort settings alone move a LiveBench score by more than that.

Published benchmarks rank models on someone else's tasks. Before committing, see LLM & agent evaluation for building an eval on your own.