// head_to_head
DeepSeek V4 Pro 0813 vs Step 5 Preview
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.
DeepSeek V4 Pro 0813 is the cheaper of the two; neither can be ranked on quality here.
DeepSeek V4 Pro 0813 does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.
deepseek
DeepSeek V4 Pro 0813
- Blended / 1M
- $1.25
- Context
- 1.0M
- Released
- Aug 12, 2026
- Overall score
- Not evaluated
stepfun
Step 5 Preview
- Blended / 1M
- $1.43
- Context
- 1.0M
- Released
- Oct 8, 2026
- Overall score
- 58.6
Specs and pricing
| Metric | DeepSeek V4 Pro 0813 | Step 5 Preview |
|---|---|---|
| LiveBench overall Weighted mean of the seven LiveBench category scores, 0–100. Higher is better. | — | 58.6 |
| Cost per point Measured benchmark spend divided by overall score — dollars per point of capability. | — | $0.0124 |
| Blended price / 1M 3:1 input:output mix, the usual shape of production traffic. | $1.25win | $1.43 |
| Input price / 1M | $0.117win | $1.00 |
| Output price / 1M | $4.66 | $2.70win |
| Cached input / 1M Price of an input token served from the prompt cache, where the provider publishes one. | $0.111 | $0.100win |
| Context window | 1.0M | 1.0M |
| Max output tokens | 944Kwin | 64K |
Benchmarks by category
An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — DeepSeek V4 Pro 0813 on top, Step 5 Preview below, both out of 100.
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.
| Workload | DeepSeek V4 Pro 0813 | Step 5 Preview |
|---|---|---|
| Support chatbot 1.2K in / 400 out × 200K requests | $400.36/mo | $391.20/mo |
| RAG assistant 8K in / 600 out × 100K requests | $370.60/mo | $602.00/mo |
| Coding agent 40K in / 4K out × 20K requests | $462.92/mo | $512.00/mo |
| Document extraction 20K in / 1.5K out × 50K requests | $465.72/mo | $1,158/mo |
| Bulk classification 500 in / 20 out × 5M requests | $754.50/mo | $2,320/mo |
Which should you pick?
You are cost-constrained
DeepSeek V4 Pro 0813
Cheaper on blended list price at $1.25 per million tokens.
DeepSeek V4 Pro 0813 vs Step 5 Preview FAQ
Which is better, DeepSeek V4 Pro 0813 or Step 5 Preview?
DeepSeek V4 Pro 0813 is the cheaper of the two; neither can be ranked on quality here. DeepSeek V4 Pro 0813 does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.
Is DeepSeek V4 Pro 0813 cheaper than Step 5 Preview?
DeepSeek V4 Pro 0813 is cheaper. On a 3:1 input:output blend, DeepSeek V4 Pro 0813 lists at $1.25 per million tokens and Step 5 Preview at $1.43 — DeepSeek V4 Pro 0813 is 14% cheaper. Input and output are priced separately — DeepSeek V4 Pro 0813 charges $0.117 in and $4.66 out, Step 5 Preview charges $1.00 and $2.70 — so the model that looks cheaper flips depending on how output-heavy your workload is.
Does DeepSeek V4 Pro 0813 or Step 5 Preview have a bigger context window?
They are effectively the same — 1.0M for DeepSeek V4 Pro 0813 and 1.0M for Step 5 Preview.
Do DeepSeek V4 Pro 0813 and Step 5 Preview support prompt caching?
Both publish a cached-input rate: $0.111 per million for DeepSeek V4 Pro 0813 and $0.100 for Step 5 Preview, against full input rates of $0.117 and $1.00. 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.
- Scores — LiveBench release 2026-10-25, 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.