Skip to main content
MetaApr 6, 2025

Llama 3.3 70B

llama-3.3-70bPrivateOpen weights

Llama 3.3 70B is Meta’s instruction-tuned 70-billion-parameter Llama model, which Meta reports performs close to its much larger Llama 3.1 405B on many tasks. It supports text input and output with a 128K-token context window, for chat, writing and code.

$0.70/$2.80
per 1M in / out
128K
context window
4K
max output
How it comparesamong 143 text models
GPQA Diamondhigher than 10%50.5%
Blended pricecheaper than 61%$1.22
Context windowlarger than 2%128K
Max outputlarger than 0%4K

Further right is better. Hover a dot for its model.

Time to first token— 
Output speed— 
Uptime— 

Last 24 hours of Venice traffic.How it’s measured

Capabilities

What the model takes in, returns and supports through the API.

InText
OutText
Streamingstream: true
Function callingtools
Web searchenable_web_search

Not supportedstructured outputs, reasoning, image input, video input, audio input, X search, prompt caching and log probabilities

Benchmarks

GPQA Diamond
50.5%122nd of 139 · 42 score higher for less, including DeepSeek v4.1 Flash (1017), MiniMax M2.5 and DeepSeek V4.1 Flash 1017 Uncensored

Specs

Model ID
llama-3.3-70b
Endpoint
POST /api/v1/chat/completions
Context window
128,000 tokens
Max output
4,096 tokens
Reasoning
None
Precision
FP8
Privacy
Private
Added
Apr 6, 2025

Compare

ModelInputOutputGPQA DiamondContextMax outputFirst tokenSpeed
Llama 3.3 70B$0.70$2.8050.5%128K4K——
Qwen 3 235B A22B Thinking 2507$0.45$3.5081.3%128K16K——
GLM 4.7$0.55$2.6554.5%198K16K——
Gemini 3 Flash Preview$0.70$3.7587.4%256K64K——
Qwen 3 Next 80b$0.35$1.9047.5%256K16K——

Other versions

  1. Llama 3.2 3BOct 2024
  2. Llama 3.3 70BApr 2025

Pricing

Billed per token in USD. Reasoning tokens bill as output.

Input$0.70per 1M tokens
Output$2.80per 1M tokens

Estimate a workload

$0.0028per request
$28.00for 10,000 requests

Blended price against similar models

Llama 3.3 70B$1.22
GLM 4.7$1.07

Per 1M tokens at three input tokens for each output token.

Quickstart

OpenAI-compatible: point any OpenAI SDK at the Venice base URL with a Venice API key.

curl https://api.venice.ai/api/v1/chat/completions \
-H "Authorization: Bearer $VENICE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "llama-3.3-70b",
"messages": [{ "role": "user", "content": "Explain TEE attestation in two sentences." }]
}'
Base URL https://api.venice.ai/api/v1

Parameters

ParameterValuesNotes
modelllama-3.3-70bRequired
messagestextContent parts the model reads
max_completion_tokensUp to 4,096
toolsFunction definitionsWith tool_choice
streamtrue | falseServer-sent events
venice_parameters.enable_web_searchauto | on | offWeb results and citations