GPT-3.5 Turbo 0125 is an OpenAI GPT-3.5 Turbo model described as a legacy GPT model for cheaper chat and non-chat tasks. OpenAI’s model documentation lists it as optimized for chat using the Chat Completions API, while also working well for non-chat tasks.
The model is presented as a 16k-context GPT-3.5 Turbo variant with improved instruction following and more reliable formatted outputs.
| Developer | Open AI |
|---|---|
| Model ID | openai/gpt-3.5-turbo-0125 |
| Release Date | August 6, 2025 |
| Context Window | 16,000 tokens (~16K) |
| Modalities | Text → Text |
| Capabilities | parallel tool calls, streaming, structured (JSON schema) output, tool/function calling |
Model positioning
OpenAI describes GPT-3.5 Turbo as a legacy GPT model for cheaper chat and non-chat tasks.
Context and output limits
OpenAI’s model page lists a 16,385-token context window and 4,096 max output tokens.
Output formatting
OpenRouter describes the model as having improved instruction following, JSON mode, reproducible outputs, and parallel function calling.
Structured responses
Tokonomix says the model’s instruction-following improvements make it more reliable for structured outputs such as JSON schemas, CSV formatting, and template-driven drafts.
Language and function-call handling
A comparison source says this version fixed a bug that caused text-encoding issues for non-English language function calls.
Customer-service automation
Tokonomix identifies customer-service automation as one of the model’s core strengths.
Lightweight code generation
Tokonomix lists lightweight code generation as a strong use case for the model.
Short-form creative writing
Tokonomix says the model performs well for short-form creative writing.
Data extraction from semi-structured text
Tokonomix says it works well for parsing invoices, extracting named entities, and tabulating product-review features when field definitions are explicit.
GPT-3.5 Turbo
A comparison source states that GPT-3.5 Turbo and GPT-3.5 Turbo 0125 share the same 16.4K input context window and 4,096 maximum output tokens.
GPT-4o
A comparison source notes that GPT-3.5 Turbo 0125 does not support image processing, unlike GPT-4o.
GPT-4 32K
A comparison source also notes that GPT-3.5 Turbo 0125 does not support image processing, unlike GPT-4 32K.
Sources: https://tokonomix.ai/en/models/openai/gpt-3.5-turbo-0125 · https://developers.openai.com/api/docs/models/gpt-3.5-turbo · https://openrouter.ai/openai/gpt-3.5-turbo-0125
Measured on the AI/ML API gateway across live traffic. Figures are medians unless stated. Last updated 2026-10-12.
GPT-3.5 Turbo is documented as suitable for cheaper chat and non-chat tasks. It is also optimized for chat using the Chat Completions API.
It can understand and generate natural language or code, which makes it suitable for language and coding tasks. It is also described as working well for non-chat tasks, including traditional completion-style use cases.
The model supports more reliable JSON outputs and is described as having improved instruction following, so it is suited to structured-output workflows. It also supports parallel function calling, making it suitable for tool-using applications.
import requests
r = requests.post(
"https://api.aimlapi.com/v1/chat/completions",
headers={"Authorization": "Bearer " + AIMLAPI_KEY},
json={
"model": "openai/gpt-3.5-turbo-0125",
"messages": [
{
"role": "user",
"content": "Hello!"
}
]
},
)
print(r.json())
const r = await fetch("https://api.aimlapi.com/v1/chat/completions", { method: "POST", headers: { Authorization: `Bearer ${process.env.AIMLAPI_KEY}`, "Content-Type": "application/json", }, body: JSON.stringify({ "model": "openai/gpt-3.5-turbo-0125", "messages": [ { "role": "user", "content": "Hello!" } ] }), }); console.log(await r.json());
curl -X POST https://api.aimlapi.com/v1/chat/completions \ -H "Authorization: Bearer $AIMLAPI_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"openai/gpt-3.5-turbo-0125","messages":[{"role":"user","content":"Hello!"}]}'
OpenAI-compatible — swap the base URL and it works with your existing SDK.
| Type | Price |
|---|---|
| Input | |
| Output | |
GPT-3.5 Turbo supports rate limits that vary by usage tier, with Free listed as not supported and higher tiers capped by RPM, TPM, and batch queue limits.
| Benchmark | Score | What it measures | Source | Retrieved |
|---|---|---|---|---|
| Intelligence | 5.5 | Composite score across standardised reasoning, knowledge and problem-solving evaluations, measured independently by Artificial Analysis | Source | September 12, 2026 |
| Coding | 10.7 | Composite score across standardised coding evaluations, measured independently by Artificial Analysis | Source | September 12, 2026 |
| Model | Input | Output | Context | Best for |
|---|---|---|---|---|
GPT-3.5 Turbo This page | Agentic workflows and structured output | |||
| Reasoning + agents | ||||
| Reasoning + agents |
GPT-3.5 Turbo has a 16,000 tokens context window.
GPT-3.5 Turbo takes text as input and returns text.
Use openai/gpt-3.5-turbo-0125 as the model id. Requests go to https://api.aimlapi.com/v1/chat/completions.
GPT-3.5 Turbo became available on August 7, 2025.
GPT-3.5 Turbo is priced at input $0.65 / 1M tokens, output $1.95 / 1M tokens.
Yes, GPT-3.5 Turbo can stream responses as they are generated.
GPT-3.5 Turbo was built by OpenAI.
No. It only supports text input and text output.
Yes, it supports tools, parallel tool calls, and structured output.
