meta/llama-4-maverick-17b-128e-instruct| Provider | Input, $ per 1M | Output, $ per 1M | In our data since |
|---|---|---|---|
| lambda_ai | $0.05 | $0.1 | 31 Jul 2026 |
| Abacus.AI | $0.14 | $0.59 | 31 Jul 2026 |
DeepInfra
deepinfra/meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8
|
$0.15 | $0.6 | 31 Jul 2026 |
| io.net | $0.15 | $0.6 | 31 Jul 2026 |
DeepInfra
meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8
|
$0.2 | $0.8 | 31 Jul 2026 |
| Groq | $0.2 | $0.6 | 31 Jul 2026 |
| Amazon Bedrock | $0.24 | $0.97 | 2 Aug 2026 |
| Azure Cognitive Services | $0.25 | $1 | 31 Jul 2026 |
Microsoft Foundry
llama-4-maverick-17b-128e-instruct-fp8
|
$0.25 | $1 | 31 Jul 2026 |
| novita-ai | $0.27 | $0.85 | 31 Jul 2026 |
| LLM Gateway | $0.27 | $0.85 | 31 Jul 2026 |
| Novita AI | $0.27 | $0.85 | 31 Jul 2026 |
| Together AI | $0.27 | $0.85 | 31 Jul 2026 |
| hyper | $0.284 | $0.934 | 31 Jul 2026 |
| Google Vertex AI | $0.35 | $1.15 | 31 Jul 2026 |
| SambaNova | $0.63 | $1.8 | 31 Jul 2026 |
| OCI Generative AI | $0.72 | $0.72 | 31 Jul 2026 |
Microsoft Foundry
azure_ai/Llama-4-Maverick-17B-128E-Instruct-FP8
|
$1.41 | $0.35 | 31 Jul 2026 |
Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.
A single provider appears several times if it sells this model under different identifiers and at different prices — for example, at different weight precision. The identifier is shown next to the name. Identical offers that arrived from two sources under different spellings are merged into one row.
The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.
| Task set | Result | Run conditions | Measured by |
|---|---|---|---|
| Arena Score in English | 1,302.96 1,298–1,308 | — | — |
| Arena Score, programming | 1,302.89 1,295–1,310 | — | — |
| Arena Score, mathematics | 1,299.94 1,289–1,311 | — | — |
| Arena Score, multi-turn dialogue | 1,288.1 1,280–1,296 | — | — |
| Arena Score, overall | 1,287.81 1,284–1,292 | — | — |
| Arena Score in Russian | 1,286.05 1,275–1,297 | — | — |
| Arena Score in Spanish | 1,284.52 1,262–1,307 | — | — |
| Arena Score, hard prompts | 1,281.63 1,276–1,287 | — | — |
| Arena Score, long queries | 1,281.1 1,274–1,289 | — | — |
| Arena Score, instruction following | 1,268.51 1,262–1,275 | — | — |
| Arena Score, creative writing | 1,267.09 1,259–1,276 | — | — |
| Arena Score, expert questions | 1,259.89 1,246–1,273 | — | — |
| Arena Score in French | 1,256.85 1,231–1,283 | — | — |
| Arena Score, text recognition in images | 1,143.61 1,132–1,155 | — | — |
| Arena Score, working with images | 1,142.98 1,134–1,152 | — | — |
| Arena Score, understanding diagrams | 1,133.57 1,115–1,153 | — | — |
| MATH, difficulty level five | 73.02 % | · with a tuned harness | Epoch evaluations |
| Creative writing (Lech Mazur’s evaluation) | 62 % | — | lechmazur/writing Github repository |
| GPQA Diamond — graduate-level questions | 55.98 % | · with a tuned harness | Epoch evaluations |
| GeoBench — locating a place from a photograph | 52 % | — | GeoBench leaderboard |
| Fiction.LiveBench — holding a long context | 46.2 % | — | Fiction.live leaderboard |
| WeirdML — unusual machine learning tasks | 24.47 % | — | WeirdML Leaderboard |
| Mock AIME 2024–2025 — olympiad problems | 20.48 % | · with a tuned harness | Epoch evaluations |
| Aider Polyglot — code edits in six languages | 15.6 % | · with a tuned harness | Aider LLM Leaderboards |
| SimpleBench — trick questions | 13.24 % | — | SimpleBench Leaderboard |
| ARC-AGI — generalising to unseen patterns | 4.4 % | · with a tuned harness | ARC Prize Leaderboard |
| Humanity’s Last Exam — expert-level questions | 0.92 % | — | — |
| CritPt — physics problems | 0 % | — | — |
| ARC-AGI-2 | 0 % | — | — |