Model index · Meta · updated 2026-08-16
Meta models and prices
17 models, 12 with a published price , 6 with open weights · released 2024-07-23 to 2026-08-09.
You can buy from Meta directly, or from 17 other hosts
We hold offerings sold by Meta itself as well as by independent hosts. Buying direct is not automatically the cheapest route, and it is not automatically the same product either — hosts differ on quantisation, served context window and rate limits. Publishes Llama as open weights and sells no per-token Llama inference, so every Llama price in the catalog is a host price. It does sell some of its own models directly — Muse Spark is listed at $1.25 per 1M input — so "Meta does not sell inference" is true of Llama and false of Meta, which is why this is recorded per model rather than per company.
What Meta makes
| Category | Models | Browse |
|---|---|---|
| Text | 11 | all Text models → |
| Image | 4 | all Image models → |
| Music & Sound | 1 | all Music & Sound models → |
| Video | 1 | all Video models → |
Cheapest Meta models per 1M tokens
Standard on-demand tier, cheapest host per model. Media models are priced in their own unit and listed further down rather than mixed into this table.
| Model | Blended /1M | Cheapest host | Quality | Context |
|---|---|---|---|---|
| Llama 3.2 1B Instruct | $0.006 | DeepInfra | not measured | 60k |
| Llama 3.2 3B Instruct | $0.020 | DeepInfra | not measured | 131k |
| Llama 3.1 8B Instruct | $0.025 | DeepInfra | not measured | 131k |
| Llama 4 Scout | $0.150 | OpenRouter (default route) * | 40/100 | 1.31M |
| Llama 3.3 70B Instruct | $0.155 | OpenRouter (default route) * | 40/100 | 131k |
| Llama Guard 4 12B | $0.180 | DeepInfra | not measured | 1.05M |
| Llama 4 Maverick | $0.324 | OpenRouter (default route) * | 80/100 | 1.05M |
| Llama 3.1 70B Instruct | $0.400 | OpenRouter (default route) * | 20/100 | 131k |
| Muse Glimmer 30B | $0.525 | DeepInfra | not measured | 131k |
| Muse Spark 1.2 | $2.00 | Meta | not measured | 1.05M |
| Muse Spark 1.1 | $2.00 | Meta | not measured | 1.05M |
Measured quality across Meta models
Our own probe suite at temperature 0, scored by deterministic checks — no LLM judge and no vendor claim. Measured on request, so most of the catalog is unscored; a blank is not measured, never a low score. Confidence falls when probes could not be read, which happens on models that answer in a hidden reasoning channel.
| Model | Score | Tier | Ranking probes | Confidence |
|---|---|---|---|---|
| Llama 4 Maverick | 80/100 | near-frontier | 4/5 | 1 |
| Llama 4 Scout | 40/100 | mid | 2/5 | 1 |
| Llama 3.3 70B Instruct | 40/100 | mid | 2/5 | 1 |
| Llama 3.1 70B Instruct | 20/100 | economy | 1/5 | 1 |
Open weights from Meta
Parameter counts are computed from the published weight files, and licences are what the authors attached — both from the weights registry rather than from any seller. Licence class decides whether you can self-host commercially without signing something.
| Model | Parameters | Licence | Downloads (30d) | Repo |
|---|---|---|---|---|
| Llama 3.2 1B Instruct | 1.24B | llama3.2 | 9.20M | meta-llama/Llama-3.2-1B-Instruct |
| Llama 3.1 8B Instruct | 8.03B | llama3.1 | 7.35M | meta-llama/Llama-3.1-8B-Instruct |
| Llama 3.2 3B Instruct | 3.21B | llama3.2 | 1.17M | meta-llama/Llama-3.2-3B-Instruct |
| Llama 3.1 70B Instruct | 70.55B | llama3.1 | 0.76M | meta-llama/Llama-3.1-70B-Instruct |
| Llama 3.3 70B Instruct | 70.55B | llama3.3 | 0.33M | meta-llama/Llama-3.3-70B-Instruct |
| Llama Guard 4 12B | 12B | llama4 | 0.20M | meta-llama/Llama-Guard-4-12B |
Who sells Meta models
Offering rows per seller across this vendor's models. A seller appearing many times serves many of the models, not the same one repeatedly.
| Seller | Offerings | |
|---|---|---|
| DeepInfra | 14 | |
| OpenRouter (default route) | 11 | |
| Novita | 7 | |
| 4 | ||
| Cloudflare | 4 | |
| Groq | 3 | |
| Parasail | 3 | |
| CoreWeave | 3 | |
| Together | 3 | |
| Meta | 2 | the vendor itself |
| Nebius | 1 | |
| AkashML | 1 | |
| Crusoe | 1 | |
| SambaNova | 1 | |
| DigitalOcean | 1 | |
| Amazon (Bedrock) | 1 | |
| Fireworks | 1 | |
| fal | 1 |
Jobs these models do
Comparison on each model page is restricted to models doing the same job — a price difference between two different jobs is arithmetic rather than advice.
| Job | Models |
|---|---|
| text generation | 11 |
| image understanding | 4 |
| text to speech | 1 |
| unclassified | 1 |
One API for every Meta model
VernaOne fronts the sellers on this page behind a single interface, so switching model or host is a config change — with automatic fallback if quality or latency regresses.
Try VernaOne free →All 17 Meta models
Newest first, priced or not. A model with no price is a fact about what its sellers publish, not a gap we are hiding.
| Model | Category | Released | Price | Kind | Job |
|---|---|---|---|---|---|
| Muse Glimmer 30B | Text | 2026-08-09 | $0.525 /1M | reasoning | text generation |
| Muse Spark 1.2 | Text | 2026-08-05 | $2.00 /1M | reasoning | text generation |
| Muse Spark 1.1 | Text | 2026-07-16 | $2.00 /1M | reasoning | text generation |
| Segment Anything Model 3 | Image | 2025-11-20 | $0.005 /request | — | image understanding |
| Demucs | Music & Sound | 2025-10-27 | no published price | — | text to speech |
| Llama Guard 4 12B | Text | 2025-04-30 | $0.180 /1M | standard open weights | text generation |
| Llama 4 Scout | Text | 2025-04-05 | $0.150 /1M | standard | text generation |
| Llama 4 Maverick | Text | 2025-04-05 | $0.324 /1M | standard | text generation |
| EVF-SAM2 Segmentation | Image | 2025-02-26 | no published price | — | image understanding |
| Segment Anything Model 2 (auto segment) | Image | 2025-02-25 | no published price | — | image understanding |
| Llama 3.3 70B Instruct | Text | 2024-12-06 | $0.155 /1M | standard open weights | text generation |
| Llama 3.2 1B Instruct | Text | 2024-09-25 | $0.006 /1M | standard open weights | text generation |
| Llama 3.2 3B Instruct | Text | 2024-09-25 | $0.020 /1M | standard open weights | text generation |
| Segment Anything Model 2 (image) | Image | 2024-08-15 | no published price | — | image understanding |
| Segment Anything Model 2 (video) | Video | 2024-08-15 | no published price | — | — |
| Llama 3.1 70B Instruct | Text | 2024-07-23 | $0.400 /1M | standard open weights | text generation |
| Llama 3.1 8B Instruct | Text | 2024-07-23 | $0.025 /1M | standard open weights | text generation |
Other companies
Alibaba · Black Forest Labs · LTX · Kling · OpenAI · Google · Stability AI · Minimax · Bria AI · Bytedance · Pixverse · Ideogram · Tencent · Luma AI · All models
Frequently asked
How many models does Meta have?
We track 17 Meta models across 4 categories, 12 of them with a published price. The list is rebuilt from provider APIs, the vendor's own docs and the weights registry rather than maintained by hand, so new releases appear without us writing about them.
Can I buy Meta models directly from Meta?
Yes — we hold offerings sold by Meta itself (as meta), alongside 17 other hosts. Buying direct is not automatically cheapest: each model page lists every seller with its price and the evidence behind it.
Are Meta models open weights?
6 of 17 have a public weights repository. Licence mix: 5 open weights restricted, 1 source available. Licence class decides whether you can self-host commercially without a separate agreement, so check the model page before planning a deployment.