Skip to main content

Llama 4 Maverick

Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forward pass (400B total). It supports multilingual text and image input, and produces multilingual text and code output across 12 supported languages. Optimized for vision-language tasks, Maverick is instruction-tuned for assistant-like behavior, image reasoning, and general-purpose multimodal interaction. Maverick features early fusion for native multimodality and a 1 million token context window. It was trained on a curated mixture of public, licensed, and Meta-platform data, covering ~22 trillion tokens, with a knowledge cutoff in August 2024. Released on April 5, 2025 under the Llama 4 Community License, Maverick is suited for research and commercial applications requiring advanced multimodal understanding and high model throughput.

Input / output price
$0.20 / $0.80 per 1M
Context
1M tokens
Released
Apr 5, 2025
Agent setup

Paste into Claude Code, Cursor, or Codex

Help me set up the Respan gateway: fetch and follow https://www.respan.ai/docs/agent-setup.md. Use model "llama-4-maverick".

Read docs

Providers

2 providers, cheapest first, OpenRouter last, priced per 1M tokens after gateway discounts.

ProviderInputOutput
Concentrateconcentrate/llama-4-maverick
$0.20/M$0.80/M
OpenRouteropenrouter/llama-4-maverick
$0.35/M$1.00/M

Is a provider down? OpenRouter status

Pricing

Averaged across 2 providers, in USD per 1M tokens after gateway discounts.

Average input
$0.28 per 1M
Average output
$0.90 per 1M
Average cache read
$0.17 per 1M

Cost by volume

Total cost from 0 to 1T tokens on the cheapest and most expensive provider.

Highest$0.35/MOpenRouterLowest$0.20/MConcentrate

Throughput and latency

30-day averages from requests through the Respan gateway. Under each figure, every model on the gateway by its average, this one in blue.

Output speed

Higher is better

21tokens/sacross providers

Faster than 50% of the other 596 models on the gateway

Response latency

Lower is better

1.0sacross providers

Faster than 62% of the other 766 models on the gateway

Benchmarks

The Respan Index plus 9 published scores from 5 sources, each linked to where it was published.

14.8/ 100

Respan Index#163 of 176

176 ranked models. The shaded band is this model's 95% interval, 10.7 to 25.6. Granite 4.2 8B and o1-mini fall inside it, so the gap is within the margin of error.
Llama 4 Maverick by category: Coding 1.9, Agents 10.2 (estimated), Reasoning 20.4, Knowledge & language 12.3, Human preference 36.5
  1. 163Llama 4 Maverick14.8
  2. 161Granite 4.2 8B14.9
  3. 164o1-mini13.9

Published scores

BenchmarkScore
LiveCodeBenchhuggingface.co43.4%
WeirdMLHåvard Tveit Ihle24.5%

Compare Llama 4 Maverick: Llama 4 Maverick vs Muse Wonder Full, Llama 4 Maverick vs Muse Wonder, Llama 4 Maverick vs Muse Ember, Llama 4 Maverick vs Muse Spark 1.3, Llama 4 Maverick vs Muse Spark 1.1, Llama 4 Maverick vs Muse Spark 1.2

Apps

Use Llama 4 Maverick in the tools you already run. Each guide points the app at the Respan gateway.

Setup guides for more apps are in the integration docs.

Llama 4 Maverick FAQ

Short answers, built from the data on this page.

Llama 4 Maverick is a model from Meta, released on Apr 5, 2025. Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forward pass (400B total). It supports multilingual text and image input, and produces multilingual text and code output across 12 supported languages. Optimized for vision-language tasks, Maverick is instruction-tuned for assistant-like behavior, image reasoning, and general-purpose multimodal interaction. Maverick features early fusion for native multimodality and a 1 million token context window. It was trained on a curated mixture of public, licensed, and Meta-platform data, covering ~22 trillion tokens, with a knowledge cutoff in August 2024. Released on April 5, 2025 under the Llama 4 Community License, Maverick is suited for research and commercial applications requiring advanced multimodal understanding and high model throughput.

On the Respan gateway, Llama 4 Maverick costs from $0.20 per 1M input tokens and $0.80 per 1M output tokens on Concentrate, after gateway discounts. Averaged across its 2 providers, it costs $0.28 per 1M input tokens and $0.90 per 1M output tokens.

Llama 4 Maverick is available through 2 providers: Concentrate and OpenRouter.

Llama 4 Maverick accepts up to 1M tokens of context on its largest provider route.

Point any OpenAI-compatible SDK at https://gateway.respan.ai/v1 with your Respan API key and set the model to "llama-4-maverick". To set it up with a coding agent, copy the agent setup prompt at the top of this page.

Llama 4 Maverick scores 14.8 out of 100 on the Respan Index v1.1, ranking #163 of 176 models. By category: Coding 1.9, Agents 10.2 (estimated), Reasoning 20.4, Knowledge & language 12.3, and Human preference 36.5. The Respan Index combines independent and lab-reported benchmark results into one score, using benchmark data collected by Respan.

Llama 4 Maverick has 9 published benchmark scores, including LiveCodeBench 43.4%, GPQA Diamond 69.8%, and MMLU-Pro 80.5%. Each score links to where it was published; the data is collected by Respan.

Certified where it counts.

Built to meet the security and privacy standards that enterprise and healthcare teams require.

ISO 27001

The internationally recognized standard for information security management.

SOC 2

Secure, compliant management of your data across all of our systems.

GDPR

Operated under GDPR, the world's strictest standard for data privacy.

HIPAA

HIPAA compliant, with a BAA available for healthcare teams.

Works with your entire stack

Use Respan with your favorite frameworks and tools.

AnthropicAI21 LabsxAIVercel AI SDK
LangChainOpenRouterAssemblyAICursor
LiteLLMGroqBasetenMem0
CohereCogneeLlamaIndexFireworks
DeepSeekLinkupHaystackTogether AI
PerplexityInference.netPostHogRubyLLM
SlackMastraAzure OpenAIMistral
AWS BedrockMoonshotHyperspellBAML
ModaInstructorGoogle GeminiParasail
ReductoZapierAgnoNebius
OpenAINovita AIReplicateSuperagent

Built for AI agents.
Break less.
Ship more.