Llama 4 Maverick
Meta · released Apr 1, 2025 · meta-llama/Llama-4-Maverick-17B-128E-Instruct
- Type
- Open weightsCustom licence
- Params
- 402B
- Context
- 1M
about 786K words of context · download allowed, licence restricts use
Our take
Written Sep 29, 2026Llama 4 Maverick is a downloadable model you can run yourself, with text and image input and room for long documents in one request. Its measured quality sits in the bottom quarter of every board we hold, so it is a model to trial on work you can check, not a safe default.
Consider it only for experiments where a downloadable model with a custom licence is acceptable and you can judge the output yourself. The licence puts conditions on commercial use and redistribution, so it needs reading before you build on it. Skip it if you need fixes landed in an existing codebase without review, or if you need measured quality that stands near the top of a board.
The case for it
- You can download it and run it yourself, so a host is optional rather than the only route.
- The request capacity takes long documents without splitting them up first, though reliable recall across all of it is unverified in our data.
- Text and images go into the same request, so a screenshot does not have to be described in words first.
The case against it
- Weak across every board we hold: 134th of 168 on Arena Text (overall) as of 25 Sep 2026, with its Creative Writing and Coding placings also in the bottom quarter of their fields.
- Coding on real issues is near the bottom of its field: 40th of 42 on SWE-bench Verified via mini-SWE-agent as of 19 Feb 2026, which measures real GitHub issues resolved end-to-end inside that harness.
- The custom licence puts conditions on commercial use and redistribution, so it needs reading before you build on it.
How good is it?
An open text model for chat and code, though it trails most models on everyday questions and coding.
- getting answers to everyday questionsArena Text (overall) · 134th of 168
- writing and completing codeArena Coding · 131st of 168
EverydayGeneral questions and everyday reasoning
Arena Text (overall)134th of 168 · 1327
CodingWriting and fixing code on its own
Arena Coding131st of 168 · 1373
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent. It is on SWE-bench Verified, in 40th of 42 with 21.
WritingDrafting and rewriting prose
Arena Creative Writing124th of 168 · 1307
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model7 scoresEvery figure we hold, from 7 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Comfortable fit
Apple M3 Ultra (80-core GPU) · 512 GB
Room to spare. 121.2 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked 4 hours ago — each listing carries its own date.
- per 1M tokens
- $0.19 in / $0.65 out
- Context served
- 1M
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouterOpenRouter's own listing | $0.19 / $0.65checked 4 hours ago | 1M | not measured | Unknown | Unknown | Unknown |
| DigitalOcean GradientThrough OpenRouter | $0.19 / $0.65checked 4 hours ago | 128K16K max reply | 7 tok/s | No | No | Confirmed |
| Novita AIfp8Direct and through OpenRouter | $0.27 / $0.85checked 4 hours ago | 1M8K max reply through OpenRouter | 34 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| Parasailfp8Through OpenRouter | $0.35 / $1.00checked 4 hours ago | 524K33K max reply | 33 tok/s | No | No | Confirmed |
| Google Vertex AIus-east5Through OpenRouter | $0.35 / $1.15checked 4 hours ago | 524K8K max reply | not measured | No | No | Confirmed |
DigitalOcean Gradient
Novita AI
Google Vertex AIAcross the 5 listings we hold: 4 say they do not train on prompts (1 of them only through OpenRouter), 0 say they do and 1 does not say. 4 appear in the zero-retention registry we check (1 of them only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| DigitalOcean GradientThrough OpenRouter | ✓ | ✓ | ✓ |
| Novita AIfp8Direct and through OpenRouter | ✗ | ✓ | ✓ |
| Parasailfp8Through OpenRouter | ✓ | ✓ | ✓ |
| Google Vertex AIus-east5Through OpenRouter | ✓ | ✓ | ✓ |
Tool calling: 4 of 5 listings say yes, 1 says no. JSON output: 5 of 5 listings say yes. Strict schema: 5 of 5 listings say yes.
When we formed this view
Recent changes
What moved
input −6% ($0.200 → $0.188 per 1M tokens), output −6% ($0.696 → $0.652 per 1M tokens)What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 5 listings does not say whether it trains on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsCustom licence, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Custom licence
This model ships custom license terms that don't map to a known template. We haven't parsed them, so commercial use, redistribution and derivatives are unverified — review the original terms before shipping.
Identifiers
- Hugging Face
- meta-llama/Llama-4-Maverick-17B-128E-Instruct
- Takes in, gives back
- Text and images in, text out
- Catalogue slug
- meta-llama-llama-4-maverick