Llama 4 Maverick Pricing

Llama 4 Maverick API pricing

Llama 4 Maverick API pricing is $0.15 per million input tokens and $0.6 per million output tokens, verified November 1, 2026.

Input (per 1M tokens)

$0.15

Output (per 1M tokens)

$0.6

Context window

1,048,576 tokens

Max output

16,384 tokens

Released

April 5, 2025

Capabilities

Llama 4 Maverick supports image input, tool use, long context.

Llama 4 Maverick API pricing has not changed since June 28, 2026.

Price-stable since June 28, 2026.

Frequently asked questions

Short answers to the most common questions about Llama 4 Maverick pricing, limits, and capabilities.

Notes

Mixture-of-experts (MoE) architecture, natively multimodal. Stronger general-purpose and vision workloads than Llama 4 Scout. Pricing varies by third-party provider (Together AI, Replicate, Fireworks) — $0.15/$0.60 is a typical low-end rate. [2026-07-12] Context Window corrected 256,000 → 1,048,576 per Meta official spec (ai.meta.com, llama.com) corroborated by llm-stats + OpenRouter. Prices unchanged ($0.15/$0.60). Max Output 16,384 confirmed. [2026-08-17] Verification attempt produced no clear result on Input — recheck next run. Output $0.60 confirmed across OpenRouter, pricepertoken and aicostcheck. Input now widely quoted at $0.20/M (OpenRouter and several aggregators) vs the $0.15 stored here; because Llama 4 is served by many third-party hosts at different rates and no source establishes $0.15 as stale, the value was left unchanged. If the $0.20 floor holds next run, raise Input $0.15 → $0.20. [2026-09-28] Verification attempt produced no clear result — recheck next run. [2026-10-09] Verification attempt produced no clear result — recheck next run. Together-specific quotes conflict ($0.27/$0.85 FP8 per llmreference vs $0.55/$2.19 per anotherwrapper); one source cites $0.15 input as the cheapest (OpenRouter) rate. Prices LEFT at $0.15/$0.60. Row's Pricing Page URL is together.ai but stored values look like cheapest-host rates — convention worth fixing. [2026-10-10] Verification attempt produced no clear result — recheck next run. Five different quotes this run: $0.17/$0.85 (anotherwrapper, attributed to Meta), $0.20/$0.60 (pricepertoken, Vercel cheapest), $0.15/$0.60 (usagepricing, aipricing.guru), $0.22/$0.85 (kunalganglani), $0.27/$0.85 (losclouds). Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-10-12] Verification attempt produced no clear result — recheck next run. Quotes still split: $0.17/$0.85 (anotherwrapper, Aug 25), $0.188/$0.652 (OpenRouter-cheapest tracker), $0.27/$0.85 (Together), $0.15/$0.60 (DeepInfra/Fireworks, March guide). One source says Meta hasn't locked GA rates. Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-10-13] Verification attempt produced no clear result — recheck next run. Together-specific quotes still split: $0.27/$0.85 (llmreference, losclouds; verified Apr 2026) vs $0.55/$2.19 (anotherwrapper comparison, July guide marked 'verify'). No source this run confirms $0.15/$0.60 as Together's rate. Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-10-15] Verification attempt produced no clear result — recheck next run. Meta publishes no first-party rate; host quotes still split: $0.15/$0.60 (DeepInfra/Fireworks, March), $0.27/$0.85 (Together), $0.20/$0.696 and $0.188/$0.652 (OpenRouter). Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-10-16] Verification attempt produced no clear result — recheck next run. Together quotes still split ($0.27/$0.85 on two comparison sites vs ~$0.55/$2.19 in a July guide); $0.15/$0.60 remains the DeepInfra rate. Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-10-19] Verification attempt produced no clear result — recheck next run. A together.ai/pricing snippet (undated) lists Llama 4 Maverick at $0.27 input / $0.85 output, matching the earlier Together quotes; DeepInfra figures across aggregators now conflict ($0.15/$0.60 vs $0.2975/$0.9775), Novita $0.28/$0.92. Prices LEFT at $0.15/$0.60. Host/convention decision still needed — Together rate would be $0.27/$0.85. [2026-10-20] Verification attempt produced no clear result — recheck next run. Still split by host: $0.15/$0.60 (DeepInfra/Fireworks per tokencost.app), $0.27/$0.85 (Together per aicostcheck), $0.17/$0.85 (anotherwrapper, attributed to Meta). Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-10-25] Verification attempt produced no clear result — recheck next run. Together quotes still split ($0.55/$2.19 anotherwrapper vs $0.27/$0.85 in an older 2025 guide); $0.15/$0.60 again attributed to DeepInfra. No Meta first-party rate. Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-10-27] Verification attempt produced no clear result — recheck next run. $0.17/$0.85 (anotherwrapper, attributed to Meta, Aug/Oct pages) and $0.188/$0.652 (computeprices, OpenRouter cheapest). $0.15/$0.60 not reproduced; no first-party Meta rate. Prices LEFT at $0.15/$0.60. Host/convention decision still needed. [2026-11-01] Verification attempt produced no clear result — recheck next run. Quotes still split: $0.27/$0.85 (vantaige, aicostcheck), $0.17/$0.85 (anotherwrapper, attributed to Meta), $0.20/$0.60 (May 2026 Together table), $0.15/$0.60 (aipricing.guru, Together). No Meta first-party rate. Prices LEFT at $0.15/$0.60. Host/convention decision still needed.