ZMIME
Meta · Released Dec 6, 2024

Llama-3.3-70B-Instruct

Popular open Llama workhorse for multilingual chat, coding, and self-hosting.

  • Open weights
  • Tool calling
Capability
126.2ECI · #144 of 158
Input
$0.59per 1M tokens
Output
$0.745per 1M tokens
Context
128K4K max output
Pricing

What it costs

Input
$0.59per million tokens
Output
$0.745per million tokens
Blended (3:1)
$0.629cheaper than 53% of priced models

Typical monthly bills

Monthly usageEstimated cost
Side project2M input + 0.5M output tokens$1.55
Team assistant25M input + 5M output tokens$18.48
Production app250M input + 50M output tokens$184.75

Median of 40 paid API providers on models.dev.

Capability

Independent benchmarks

126.2Epoch Capabilities Index
#144 of 158 scored models
68Median 147174

The shaded band is Epoch AI’s confidence range (123.1–129.0); the tick marks the median scored model.

  • GPQA DiamondGraduate-level science questions47.4%
  • OTIS Mock AIME 2024–2025Competition mathematics5.1%

Scores from Epoch AI, run independently of Meta.

Specifications

The details

Lab
Meta
Released
Dec 6, 2024
Knowledge cutoff
Dec 2023
Context window
128,000 tokens
Max output
4,096 tokens
Inputs
Text
Output
Text
Reasoning
No
Tool calling
Yes
Structured output
No
Availability
43 API providerslisted on models.dev
Lab-reported

What Meta claims

Published by the lab at launch. Settings vary, so compare these only with care.

BenchmarkScoreSettingSource
Artificial Analysis Coding Index10.7 index—Source
SciCode26 percent correct—Source
Terminal-Bench Hard3 success rate—Source
Same lab

More from Meta

Questions

About Llama-3.3-70B-Instruct

How much does Llama-3.3-70B-Instruct cost?

Llama-3.3-70B-Instruct costs $0.59 input / $0.745 output per million tokens (median across 40 API providers). At a 3:1 input-to-output mix that is $0.629 per million tokens, cheaper than 53% of the 469 priced models we track.

What is the context window of Llama-3.3-70B-Instruct?

Llama-3.3-70B-Instruct accepts up to 128,000 tokens per request and can write up to 4,096 tokens in one response.

How good is Llama-3.3-70B-Instruct?

Epoch AI gives Llama-3.3-70B-Instruct a Capabilities Index score of 126.2 (likely range 123.1–129.0), ranking it #144 of 158 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 47.4%, OTIS Mock AIME 2024–2025 5.1%.

Is Llama-3.3-70B-Instruct open source?

Yes. Meta publishes the weights, so it can be downloaded and self-hosted.

What inputs does Llama-3.3-70B-Instruct support?

Llama-3.3-70B-Instruct accepts text and replies in text. It is not a dedicated reasoning model, supports tool calling.

When was Llama-3.3-70B-Instruct released?

Meta released Llama-3.3-70B-Instruct on Dec 6, 2024. Its training data runs to Dec 2023.

What are the best alternatives to Llama-3.3-70B-Instruct?

The closest current models from other labs on capability, price and release date are Mistral Small 3.1 24B (Mistral AI, $0.351 / $0.555), Claude Haiku 3.5 (Anthropic, ECI 124.8, $0.80 / $4.00), DeepSeek-V3 (DeepSeek, ECI 131.7, $0.32 / $1.10) and GPT-4o mini (OpenAI, ECI 123.8, $0.15 / $0.60).