Pricing comparison

Anthropic vs Meta Llama Pricing

Per-million-token pricing for Anthropic and Meta Llama, with side-by-side flagship models, cheapest tiers, and context windows. Pricing data syncs weekly from a continuously-updated model catalog — last updated May 22, 2026.

As of May 22, 2026, Meta Llama offers the lowest output-token price at $0.02/1M (Llama 3.2 3B Instruct — $0.02/1M output).

Who wins on what

Cheapest input tokens

$0.02/1M

Meta Llama

Llama 3.1 8B Instruct — $0.02/1M input

Cheapest output tokens

$0.02/1M

Meta Llama

Llama 3.2 3B Instruct — $0.02/1M output

Longest context window

1.0M

Meta Llama

Llama 4 Maverick — 1.0M input tokens

Lowest average output cost

$0.67/1M

Meta Llama

Provider-wide average across 22 models

Largest model catalog

39 models

Anthropic

More options to match cost vs capability

Most reasoning models

1 models

Anthropic

Models with dedicated reasoning / thinking support

Most vision models

7 models

Anthropic

Models that accept image input

Side-by-side

39 models

Anthropic

Cheapest input

$0.250

Claude 3 Haiku

Cheapest output

$1.25

Claude 3 Haiku

Longest context

1.0M

Claude 4 Sonnet (2025-05-14)

Avg output / 1M

$24.88

Across catalog

Cheapest cached input

$0.030

Claude Haiku 3

ModelIn/1MOut/1MCtx
Claude Opus 4.7$5.00$25.00200K
Claude Opus 4.6$5.00$25.00200K
Claude Opus 4.5$5.00$25.00200K
Claude Haiku 4.5$1.00$5.00200K
Claude Sonnet 4.5$3.00$15.001.0M
Claude 3 Haiku$0.250$1.25200K
22 models

Meta Llama

Cheapest input

$0.020

Llama 3.1 8B Instruct

Cheapest output

$0.020

Llama 3.2 3B Instruct

Longest context

1.0M

Llama 4 Maverick

Avg output / 1M

$0.674

Across catalog

ModelIn/1MOut/1MCtx
Llama 3.1 405B (base)$4.00$4.0033K
Llama 3.1 405B Instruct$3.50$3.50131K
Llama 4 Maverick$0.150$0.6001.0M
Llama 3 70B Instruct$0.300$0.4008K
Llama 3.1 70B Instruct$0.400$0.400131K
Llama 3.2 3B Instruct$0.020$0.020131K

All prices in USD per 1 million tokens. Showing top 6 models per provider, sorted by output cost.

Related comparisons

Run the numbers for your workload

Calcaas multiplies per-token costs by your real usage patterns — inputs, outputs, retries, and conversation history — across both providers in one model.