Meta / Llama · Pricing

Llama 3.1 70B pricing

Complete Llama 3.1 70B API cost reference: $0.40/M input, $0.40/M output, 131K token context window. Verified against official sources.

Prices verified August 20, 2026
Llama 3.1 70B
Meta / Llama · official pricing ↗
Input / M
$0.40
Output / M
$0.40
Context
131K
Batch
Vision
Reasoning

What Llama 3.1 70B costs in practice

WorkloadCost / callCost @ 1K req/day
Light · 1K in / 300 out$0.0005$15.83/mo
Standard · 10K in / 2K out$0.0048$146.10/mo
Heavy · 100K in / 20K out$0.0480$1461.00/mo

Estimates assume medium reasoning effort, no batch discounts and no prompt caching. Turn on prompt caching and the input side typically drops 50–90% — run your own mix in the free calculator preloaded with Llama 3.1 70B.

About this page

Prices were last verified on August 20, 2026 against Meta / Llama's official pricing page. Providers reprice frequently — see our methodology for exactly how estimates are computed, and always confirm final rates with the provider before budgeting.

Llama 3.1 70B FAQ

How much does Llama 3.1 70B cost?

Llama 3.1 70B charges $0.40 per million input tokens and $0.40 per million output tokens.

What is the context window of Llama 3.1 70B?

Llama 3.1 70B supports a context window of 131,072 tokens (131K) — roughly 98,304 English words or about 197 pages of text in a single request.

Does Llama 3.1 70B support images, audio or reasoning?

Llama 3.1 70B offers text-only processing. Hosted via Groq / Together / other providers.

How do I estimate my Llama 3.1 70B bill?

A typical workload of 10K input and 2K output tokens costs about $0.0048 per call with Llama 3.1 70B. At 1,000 requests per day that is roughly $146.10 per month. Use the free AITokenCalculator to run your exact numbers.