// head_to_head
Kimi K3 vs GPT-4o (2024-05-13)
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.
Kimi K3 is the cheaper of the two; neither can be ranked on quality here.
GPT-4o (2024-05-13) 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.
moonshotai
Kimi K3
- Blended / 1M
- $6.00
- Context
- 1.0M
- Released
- Jul 16, 2026
- Overall score
- 79.2
openai
GPT-4o (2024-05-13)
- Blended / 1M
- $7.50
- Context
- 128K
- Released
- May 13, 2024
- Overall score
- Not evaluated
Specs and pricing
| Metric | Kimi K3 | GPT-4o (2024-05-13) |
|---|---|---|
| LiveBench overall Mean of the seven LiveBench category scores, 0–100. Higher is better. | 79.2 | — |
| Cost per point Measured benchmark spend divided by overall score — dollars per point of capability. | $0.1909 | — |
| Blended price / 1M 3:1 input:output mix, the usual shape of production traffic. | $6.00win | $7.50 |
| Input price / 1M | $3.00win | $5.00 |
| Output price / 1M | $15.00 | $15.00 |
| Cached input / 1M Price of an input token served from the prompt cache, where the provider publishes one. | $0.300 | — |
| Context window | 1.0Mwin | 128K |
| Max output tokens | 944Kwin | 4K |
Benchmarks by category
An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — Kimi K3 on top, GPT-4o (2024-05-13) 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 | Kimi K3 | GPT-4o (2024-05-13) |
|---|---|---|
| Support chatbot 1.2K in / 400 out × 200K requests | $1,726/mo | $2,400/mo |
| RAG assistant 8K in / 600 out × 100K requests | $2,220/mo | $4,900/mo |
| Coding agent 40K in / 4K out × 20K requests | $2,088/mo | $5,200/mo |
| Document extraction 20K in / 1.5K out × 50K requests | $3,990/mo | $6,125/mo |
| Bulk classification 500 in / 20 out × 5M requests | $7,650/mo | $14,000/mo |
Which should you pick?
You need to fit large documents in one call
Kimi K3
Wider context window — 1.0M against 128K.
Kimi K3 vs GPT-4o (2024-05-13) FAQ
Which is better, Kimi K3 or GPT-4o (2024-05-13)?
Kimi K3 is the cheaper of the two; neither can be ranked on quality here. GPT-4o (2024-05-13) 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 Kimi K3 cheaper than GPT-4o (2024-05-13)?
Kimi K3 is cheaper. On a 3:1 input:output blend, Kimi K3 lists at $6.00 per million tokens and GPT-4o (2024-05-13) at $7.50 — Kimi K3 is 25% cheaper. Input and output are priced separately — Kimi K3 charges $3.00 in and $15.00 out, GPT-4o (2024-05-13) charges $5.00 and $15.00 — so the model that looks cheaper flips depending on how output-heavy your workload is.
Does Kimi K3 or GPT-4o (2024-05-13) have a bigger context window?
Kimi K3 has the larger context window: 1.0M for Kimi K3 against 128K for GPT-4o (2024-05-13). Note that a window you can fill is not a window you should fill — retrieval quality usually degrades well before the limit, and you pay for every token you put in it.
Do Kimi K3 and GPT-4o (2024-05-13) support prompt caching?
Kimi K3 publishes a cached-input rate of $0.300 per million tokens against a full input rate of $3.00. The catalogue lists no separate cached rate for GPT-4o (2024-05-13), which means the provider does not price it separately here — not that caching is unavailable.
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 2026-06-25, 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.