Llama 3.3 70B Instruct API — pricing & examples

Q: Does it think before answering?

No. There is no reasoning field and no reasoning tokens; max_tokens is spent entirely on the visible answer.

Models/Chat/Llama 3.3 70B Instruct

ChatTool callingJSON modeMultilingual

Context window

131K tokens

Modality

Text → Text

Input

$0.20 / 1M tokens

Output

$0.50 / 1M tokens

Time to first token

0.62 s

Output speed

98.2 tok/s

Model ID

meta-llama/Llama-3.3-70B-Instruct-Turbo

Quickstart

You need an API key to call this model. Create a free account or log in.

Set INFERENCE_API_KEY to your key from the API Keys page, then run:

curl https://api.inferenceapis.com/v1/chat/completions \
  -H "Authorization: Bearer $INFERENCE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Hello! What can you do?"}
    ]
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.inferenceapis.com/v1",
    api_key=os.environ["INFERENCE_API_KEY"],
)
response = client.chat.completions.create(
    model="meta-llama/Llama-3.3-70B-Instruct-Turbo",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello! What can you do?"},
    ],
)
print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.inferenceapis.com/v1",
  apiKey: process.env.INFERENCE_API_KEY,
});
const response = await client.chat.completions.create({
  model: "meta-llama/Llama-3.3-70B-Instruct-Turbo",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "Hello! What can you do?" },
  ],
});
console.log(response.choices[0].message.content);

When to use it

A good fit when

Your prompts, evals or output formats were tuned on Llama 3.3 70B and you do not want to re-tune
You were on Groq: the llama-3.3-70b-versatile id is accepted unchanged

Look elsewhere when

You are starting fresh with no Llama-tuned prompts to preserve

What it costs in practice

Workload	Cost
Chat assistant — 10,000 turns of 800 tokens in, 300 out	$3.10
Long prompts or RAG — 10,000 requests of 8,000 in, 500 out	$18.50
Generation-heavy — 10,000 requests of 500 in, 2,000 out	$11.00

Computed from the live rates below. This model does not reason before answering, so output tokens are the visible answer only. There is no subscription or minimum; new accounts start with free credit.

Endpoint

POST https://api.inferenceapis.com/v1/chat/completions

Headers Authorization: Bearer $INFERENCE_API_KEY
Content-Type: application/json

Compatible OpenAI wire format. /openai/v1/… is accepted too, so Groq-style base URLs work with only the host changed.

Model IDs and aliases

Send any of these as model; they all resolve to this model and bill at its rate.

ID	Note
`meta-llama/Llama-3.3-70B-Instruct-Turbo`	Canonical
`llama-3.3-70b`	Alias
`llama-3.3-70b-instruct`	Alias
`llama-3.3-70b-versatile`	Same id as on Groq — switch by changing the base URL only
`llama-3.1-8b-instant`	We do not host the 8B model. The id is accepted so broken apps start working again, but it is answered by Llama 3.3 70B and billed at this model's rate
`meta-llama/Llama-3.3-70B-Instruct`	Alias

Request body

Parameter	Type	Description
`model` required	string	Model ID: `meta-llama/Llama-3.3-70B-Instruct-Turbo` — also accepted: `llama-3.3-70b`, `llama-3.3-70b-instruct`, `llama-3.3-70b-versatile`, `llama-3.1-8b-instant`
`messages` required	array	Conversation so far. Each item has a `role` (`system`, `user` or `assistant`) and `content`.
`max_tokens`	integer	Maximum number of tokens to generate.
`temperature`	number	Sampling temperature. Higher values give more varied output.
`top_p`	number	Nucleus sampling cutoff.
`stream`	boolean	Stream the response as server-sent events. The final chunk includes `usage`.
`tools`	array	Tool definitions for function calling (OpenAI format).
`response_format`	object	`{"type": "json_object"}` for JSON mode.
`stop`	string \| array	Sequences where the model stops generating.
`seed`	integer	Seed for more reproducible sampling.

Response

200 OK · application/json

{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hi! I can answer questions, write and edit text, help with code and more."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": { "prompt_tokens": 28, "completion_tokens": 17, "total_tokens": 45 }
}

Tested on this endpoint

Tool calling (`tools`, `tool_calls`)	Works
JSON mode (`response_format: json_object`)	Works
Schema-constrained output (`response_format: json_schema`)	Works
Streaming with `usage` in the final chunk	Works
System prompt followed	Works

Each row is a live request we sent on September 17, 2026, not a claim copied from a model card. Streamed and parallel tool calls and tool_choice were tested too: full matrix for every model. Request examples are in the API documentation.

Pricing

Input	$0.20 / 1M tokens
Output	$0.50 / 1M tokens
Billing	Pay as you go from prepaid credits. No subscription, no daily request or token cap. See all pricing.

Notes

Served as the Turbo (FP8) variant. Output is very close to the full-precision model; if you compare against Groq's former endpoint you may see minor wording differences.

Compare Llama 3.3 70B Instruct with…

GPT-OSS 120B DeepSeek V4 Flash DeepSeek V4.1 Flash GLM 5.3 Flash GLM 5.3 DeepSeek V4 Pro MiniMax M3 Kimi K2.7 Code GPT-OSS 20B Qwen3.8 Flash Qwen3-VL 235B GPT-OSS 120B on Groq Llama 3.3 70B on Groq GPT-4o mini GPT-4.1 mini Gemini 2.5 Flash DeepSeek V4 Flash (DeepSeek API)

Questions about Llama 3.3 70B Instruct

Is this the same model as Groq's llama-3.3-70b-versatile?

The same Meta weights. It is served here as the FP8 "Turbo" variant, so wording can differ slightly from Groq's former endpoint on identical prompts; behaviour, prompt format and tool calling are the same. The Groq id is accepted as is.

My app used llama-3.1-8b-instant. What happens to it here?

The id is accepted and answered by Llama 3.3 70B, so the app works again without a code change, at the 70B price. If cost matters, change the model string to deepseek-ai/DeepSeek-V4-Flash or openai/gpt-oss-120b, which are both much cheaper than the 70B.

Why does it cost more than GPT-OSS 120B?

Llama 3.3 70B is a dense model: all 70 billion parameters run for every token. GPT-OSS 120B is a mixture-of-experts model that activates about 5 billion per token, so it is cheaper to serve despite the larger name. If you are not tied to Llama, it is the better deal.

Does it think before answering?

No. There is no reasoning field and no reasoning tokens; max_tokens is spent entirely on the visible answer.

What does the licence require?

The Llama 3.3 Community License allows commercial use, asks for "Built with Llama" attribution in products that use it, and has extra terms for services above 700 million monthly users.

Licence

Llama 3.3 70B Instruct is published by Meta under the Llama 3.3 Community License. Your use of its output is also subject to that licence.

Errors

Errors use the OpenAI envelope: {"error": {"message", "type", "code"}}. 401 missing or invalid key · 402 insufficient_balance · 404 model_not_found · 503 model_unavailable · 502 backend error, safe to retry. Full table with what to retry: API documentation. Errors from other providers: provider error reference.