// head_to_head
GPT-3.5 Turbo Instruct vs GPT-5.4 Mini
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.
GPT-3.5 Turbo Instruct and GPT-5.4 Mini are priced within ~10% of each other.
GPT-3.5 Turbo Instruct 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.
openai
GPT-3.5 Turbo Instruct
- Blended / 1M
- $1.63
- Context
- 4K
- Released
- Sep 28, 2023
- Overall score
- Not evaluated
openai
GPT-5.4 Mini
- Blended / 1M
- $1.69
- Context
- 400K
- Released
- Mar 17, 2026
- Overall score
- 66.4
Specs and pricing
| Metric | GPT-3.5 Turbo Instruct | GPT-5.4 Mini |
|---|---|---|
| LiveBench overall Mean of the seven LiveBench category scores, 0–100. Higher is better. | — | 66.4 |
| Cost per point Measured benchmark spend divided by overall score — dollars per point of capability. | — | $0.1871 |
| Blended price / 1M 3:1 input:output mix, the usual shape of production traffic. | $1.63 | $1.69 |
| Input price / 1M | $1.50 | $0.750win |
| Output price / 1M | $2.00win | $4.50 |
| Cached input / 1M Price of an input token served from the prompt cache, where the provider publishes one. | — | $0.075 |
| Context window | 4K | 400Kwin |
| Max output tokens | 4K | 128Kwin |
Benchmarks by category
An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — GPT-3.5 Turbo Instruct on top, GPT-5.4 Mini 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 | GPT-3.5 Turbo Instruct | GPT-5.4 Mini |
|---|---|---|
| Support chatbot 1.2K in / 400 out × 200K requests | $520.00/mo | $491.40/mo |
| RAG assistant 8K in / 600 out × 100K requests | $1,320/mo | $600.00/mo |
| Coding agent 40K in / 4K out × 20K requests | $1,360/mo | $582.00/mo |
| Document extraction 20K in / 1.5K out × 50K requests | $1,650/mo | $1,054/mo |
| Bulk classification 500 in / 20 out × 5M requests | $3,950/mo | $1,988/mo |
Which should you pick?
You need to fit large documents in one call
GPT-5.4 Mini
Wider context window — 400K against 4K.
GPT-3.5 Turbo Instruct vs GPT-5.4 Mini FAQ
Which is better, GPT-3.5 Turbo Instruct or GPT-5.4 Mini?
GPT-3.5 Turbo Instruct and GPT-5.4 Mini are priced within ~10% of each other. GPT-3.5 Turbo Instruct 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 GPT-3.5 Turbo Instruct cheaper than GPT-5.4 Mini?
They cost about the same. Both land near $1.63 per million tokens on a 3:1 input:output blend, so price is unlikely to be the deciding factor between them.
Does GPT-3.5 Turbo Instruct or GPT-5.4 Mini have a bigger context window?
GPT-5.4 Mini has the larger context window: 4K for GPT-3.5 Turbo Instruct against 400K for GPT-5.4 Mini. 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 GPT-3.5 Turbo Instruct and GPT-5.4 Mini support prompt caching?
GPT-5.4 Mini publishes a cached-input rate of $0.075 per million tokens against a full input rate of $0.750. The catalogue lists no separate cached rate for GPT-3.5 Turbo Instruct, 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.