Meta

Meta is an AI model provider.Tokenando tracks 38 Meta models, with input pricing from $0.0200/M and an average blended cost of $0.7613/M. Its flagship model is Llama 4 Maverick.

Llama open-weights family, the most-deployed open foundation models. Run them anywhere or rent them from any inference provider.

Founded 2013HQ Menlo Park, USAWebsite ↗API docs ↗

ModelsUSD per million tokens. Blend is 70% input, 30% output, our calculation.

MODELS TRACKED
38
3 categories
FLAGSHIP
Llama 4 Maverick
text+image->text
MIN INPUT
$0.0200/M
cheapest model in family
AVG BLENDED
$0.7613/M
across 38 priced models
MAX CONTEXT
1,048,576
largest window in family

Frontier

1 model

Multimodal

14 models
Muse Glimmer 30B
text+image->text · 131,072 ctx
in $0.3500/Mout $1.50/M
tokenizer: Other · cron:openrouter
Llama Guard 4 12B
text+image->text · 163,840 ctx
in $0.1800/Mout $0.1800/M
tokenizer: Other · cron:openrouter
Llama 3.2 11B Vision InstructprofileDEPRECATED
text+image->text · 131,072 ctx
in $0.2450/Mout $0.2450/M
tokenizer: Llama3 · cron:openrouter
Llama 4 Maverickprofile
text+image->text · 1,048,576 ctx
in $0.3500/Mout $1.00/M
tokenizer: Llama4 · script:pricing-refresh
Llama 4 Scoutprofile
text+image->text · 327,680 ctx
in $0.1000/Mout $0.3000/M
tokenizer: Llama4 · cron:openrouter
Muse Spark 1.3
text+image+file+audio+video->text · 1,048,576 ctx
in $1.25/Mout $4.25/M
tokenizer: Other · cron:openrouter
Muse Ember
text+image+file+audio+video->text · 1,048,576 ctx
in $0.1000/Mout $0.5000/M
tokenizer: Other · cron:openrouter
Muse Spark 1.2 Contributor
text+image+file+audio+video->text · 1,048,576 ctx
in $0.1000/Mout $0.2000/M
tokenizer: Other · cron:openrouter
Muse Spark Open
text+image+file+audio+video->text · 1,048,576 ctx
in $1.25/Mout $4.25/M
tokenizer: Other · cron:openrouter
Muse Spark 1.2profile
text+image+file+audio+video->text · 1,048,576 ctx
in $1.25/Mout $4.25/M
tokenizer: Other · cron:openrouter
Muse Wonder
text+image+file+audio+video->text · 1,048,576 ctx
in $4.00/Mout $20.00/M
tokenizer: Other · cron:openrouter
Muse Spark 1.3 Contributor
text+image+file+audio+video->text · 1,048,576 ctx
in $0.1000/Mout $0.2000/M
tokenizer: Other · cron:openrouter
Muse Spark 1.1
text+image+file+audio+video->text · 1,048,576 ctx
in $1.25/Mout $4.25/M
tokenizer: Other · cron:openrouter
Muse Glimmer 30B (batch)
text+image->text · 131,072 ctx
in $0.2400/Mout $0.9600/M

Efficient

23 models
Llama 3.2 3B Instructprofile
text->text · 80,000 ctx
in $0.0509/Mout $0.3350/M
tokenizer: Llama3 · job:openrouter
Llama 3 8B InstructprofileDEPRECATED
text->text · 8,192 ctx
in $0.0300/Mout $0.0400/M
tokenizer: Llama3 · cron:openrouter
Llama 3.3 70B Instructprofile
text->text · 131,072 ctx
in $0.7100/Mout $0.7100/M
tokenizer: Llama3 · cron:openrouter
Llama 3 70B InstructprofileDEPRECATED
text->text · 8,192 ctx
in $0.5100/Mout $0.7400/M
tokenizer: Llama3 · cron:openrouter
Llama 3.3 70Bprofile
Via API · 128,000 ctx
in $0.2300/Mout $0.4000/M
Strong open weights · manual-seed
Llama 3.1 70Bprofile
Via API · 128,000 ctx
in $0.3500/Mout $0.4000/M
Balanced open · manual-seed
Llama 3.1 8Bprofile
Via API · 128,000 ctx
in $0.1000/Mout $0.1000/M
Compact open · manual-seed
Llama 3.2 90B Visionprofile
Via API · 128,000 ctx
in $0.8800/Mout $0.8800/M
Vision + language · manual-seed
Llama 3.2 11B Visionprofile
Via API · 128,000 ctx
in $0.1600/Mout $0.1600/M
Compact vision · manual-seed
Llama 3.2 3Bprofile
Via API · 128,000 ctx
in $0.0600/Mout $0.0600/M
Ultra-compact · manual-seed
Llama 3.2 1Bprofile
Via API · 128,000 ctx
in $0.0400/Mout $0.0400/M
Smallest Llama · manual-seed
Llama 3 70Bprofile
Via API · 8,000 ctx
in $0.5900/Mout $0.7900/M
Llama 3 base · manual-seed
Llama 3 8Bprofile
Via API · 8,000 ctx
in $0.2000/Mout $0.2000/M
Llama 3 compact · manual-seed
Llama 2 70B Chatprofile
Via API · 4,000 ctx
in $0.9000/Mout $0.9000/M
Classic Llama 2 · manual-seed
Llama 2 13B Chatprofile
Via API · 4,000 ctx
in $0.2200/Mout $0.2200/M
Mid Llama 2 · manual-seed
Llama 2 7B Chatprofile
Via API · 4,000 ctx
in $0.1800/Mout $0.1800/M
Compact Llama 2 · manual-seed
CodeLlama 70B Instructprofile
Via API · 16,000 ctx
in $0.9000/Mout $0.9000/M
Largest CodeLlama · manual-seed
CodeLlama 34B Instructprofile
Via API · 16,000 ctx
in $0.3500/Mout $0.3500/M
Strong code model · manual-seed
CodeLlama 13B Instructprofile
Via API · 16,000 ctx
in $0.2200/Mout $0.2200/M
Mid code model · manual-seed
CodeLlama 7B Instructprofile
Via API · 16,000 ctx
in $0.1800/Mout $0.1800/M
Compact code · manual-seed
Llama 3.2 1B Instructprofile
text->text · 60,000 ctx
in $0.0270/Mout $0.2010/M
tokenizer: Llama3 · cron:openrouter
Llama 3.1 8B Instructprofile
text->text · 16,384 ctx
in $0.0200/Mout $0.0400/M
tokenizer: Llama3 · script:pricing-refresh
Llama 3.1 70B Instructprofile
text->text · 131,072 ctx
in $0.7200/Mout $0.7200/M
tokenizer: Llama3 · script:pricing-refresh

Frequently Asked Questions

How many models does Meta offer?

Tokenando tracks 38 Meta models.

How much do Meta models cost?

Meta model input pricing starts at $0.0200 per million tokens, with an average blended cost of $0.7613 per million across the 38 priced models we track.

What is Meta's flagship model?

Meta's flagship model is Llama 4 Maverick. It is the highest-tier Meta model we track, with input pricing of $0.3500 per million tokens.

What model categories does Meta cover?

Meta covers 3 categories: frontier, multimodal and efficient.

Compared with

Terms used on this page