Meta-llama

Llama 3.3 70B Instruct API Pricing 2026

Compare pricing, benchmarks, and providers for Llama 3.3 70B Instruct. Find the best value for your use case.

8 Ways to Use Fewer Tokens
Last updated: November 21, 2026 at 08:43 AM

Overview

Llama 3.3 70B Instruct was released on December 6, 2024. Pricing starts at $0.100 per million input tokens and $0.320 per million output tokens. The model supports a context window of up to 131K tokens. API access is available through Meta-llama.

Context Window
131K
tokens
Pricing
Input$0.100
Output$0.320
Cached$0.100
per 1M tokens
Speed
Output-
TTFT-
median latency
Capabilities
Tools

Llama 3.3 70B Instruct is priced at $0.100 per million input tokens and $0.320 per million output tokens. It has a context window of 131,072 tokens. It is available from 14 providers, with DeepInfra the cheapest at $0.100 per million input tokens (input prices range from $0.100 to $1.04). On benchmarks, Llama 3.3 70B Instruct scores Intelligence 44.9 (87th percentile), MMLU Pro 48.1 (97th percentile), GPQA 10.5 (33rd percentile), Aider 59.4 (48th percentile).

Pricing Comparison

Compare Llama 3.3 70B Instruct with 10 similar models by price.

Current Pricing (per 1M tokens)

11 models

Sponsored The Grid | Spot Priced LLM API
The Grid | Spot Priced LLM API — Why pay list price when providers will bid for your API usage? Up to 80% off.
View spot prices
Provider
Model
Input $/M
Output $/M
Coding
MMLU
GPQA
Context
Actions
$0.100
$0.320
—
48.1
10.5
131,072
Try
$0.100
$0.100
—
—
—
131,072
Try
$0.100
$0.350
—
—
—
-
Try
$0.100
$0.100
—
—
—
131,072
Try
$0.100
$0.100
—
—
—
131,072
Try
$0.100
$0.100
—
—
—
128,000
Try
$0.100
$0.400
—
—
—
1,048,576
Try
$0.100
$0.000
—
—
—
-
Try
$0.100
$0.400
—
—
—
1,047,576
Try
$0.100
$0.300
—
—
—
327,680
Try
$0.100
$0.300
—
—
—
1,048,576
Try

* Some models use tiered pricing based on prompt length. Displayed prices are for prompts ≤ 200k tokens.

Compare Providers

Llama 3.3 70B Instruct is available from multiple providers with different pricing and availability.

14
Providers
$0.100
Cheapest Input/M
$0.320
Cheapest Output/M
DeepInfra
Best Price

Provider Pricing

14 models

Sponsored The Grid | Spot Priced LLM API
The Grid | Spot Priced LLM API — Why pay list price when providers will bid for your API usage? Up to 80% off.
View spot prices
Provider
Input $/M
Output $/M
Cache Read $/M
Cache Write $/M
Uptime
DeepInfra
DeepInfra
Best Price
$0.100
$0.320
-
-
100.0%
A
AkashML
$0.200
$0.520
$0.100
-
100.0%
$0.220
$0.500
$0.110
-
100.0%
$0.293
$2.253
-
-
100.0%
$0.390
$0.390
-
-
-
A
Avian
$0.450
$0.450
-
-
-
$0.590
$0.790
$0.295
-
100.0%
C
CoreWeave
$0.710
$0.710
$0.710
-
100.0%
$0.710
$0.710
$0.710
-
-
V
Vercel
$0.720
$0.720
-
-
-
$0.720
$0.720
-
-
100.0%
$0.900
$0.900
$0.450
-
-
$1.040
$1.040
-
-
-
Llama
Llama
$N/A
$N/A
-
-
-
Provider pricing data sourced from OpenRouter and Helicone

Community Rankings

How does Llama 3.3 70B Instruct perform? Vote based on your experience.

Cost vs. Quality

Compare Llama 3.3 70B Instruct's benchmark performance against all models.

X-axis:
Y-axis:
Loading chart...
Llama 3.3 70B Instruct
Other models

Detailed Benchmark Scores

Intelligence
44.9 (87th pct)
Coding
-
Math
-
MMLU Pro
48.1 (97th pct)
GPQA
10.5 (33rd pct)
LiveCodeBench
-
AIME
-
Aider
59.4 (48th pct)
Benchmark data from Artificial Analysis and HuggingFace Open LLM Leaderboard

Try Llama 3.3 70B Instruct

Use our Calculator

Estimate your costs based on expected token usage.

Open Cost Calculator

Try in Playground

Test Llama 3.3 70B Instruct directly in your browser.

Open Playground

Frequently Asked Questions

Llama 3.3 70B Instruct costs $0.000100 per 1,000 input tokens and $0.000320 per 1,000 output tokens.
Llama 3.3 70B Instruct is available from 14 providers. DeepInfra offers the best price.
DeepInfra offers Llama 3.3 70B Instruct at $0.100 per million input tokens, the lowest of 14 providers (input prices range from $0.100 to $1.04 per million tokens).
Llama 3.3 70B Instruct scores Intelligence 44.9 (87th percentile), MMLU Pro 48.1 (97th percentile), GPQA 10.5 (33rd percentile), Aider 59.4 (48th percentile). Only benchmarks with published data for this model are shown.
We compare Llama 3.3 70B Instruct with 10 similarly-priced models. See the benchmark scatter plot above to compare quality vs cost.

All Meta-llama Models

See pricing for all Meta-llama models.

View all Meta-llama models →