Skip to content
OfferSign up and get $10 of credit for any model in Chat and the API. Top up from $1, with no processing fees. See free models →·Some free models are at the daily cap — refills 05:30 IST; requests auto-route to a live backup. Status →
One AI API Key
Get API key

Meta · Muse · Balanced

Muse Spark Open

Discount −30%New

Meta's open-weight model for coding and agent workflows, with text and image input.

Model id

muse-spark-open
Get an API keyTry in Chat

Price per 1M tokens

Input
Regular rate $1.25now $0.875
Output
Regular rate $4.25now $2.975
Cached input
$0.1583% below input

30% below the market rate.

Added here 18 Oct 2026 — read the changelog for what changed and why.

Output is 3.4× the input rate. $1 covers about 423 prompts of 1,000 input + 500 output tokens. 1 credit = $1 of usage. All rates

Start using Muse Spark Open

Set ONEAIAPIKEY_API_KEY to a key from your console, then run any example. Streaming is on so long answers are not cut off by idle-connection timeouts.

curl https://api.oneaiapikey.com/v1/chat/completions \
  -H "Authorization: Bearer $ONEAIAPIKEY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "muse-spark-open",
    "stream": true,
    "messages": [{"role": "user", "content": "Write a TypeScript endpoint that returns paginated orders for the signed-in user."}]
  }'
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.oneaiapikey.com/v1",
    api_key=os.environ["ONEAIAPIKEY_API_KEY"],
)

stream = client.chat.completions.create(
    model="muse-spark-open",
    messages=[{"role": "user", "content": "Write a TypeScript endpoint that returns paginated orders for the signed-in user."}],
    stream=True,  # recommended for long outputs
)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.oneaiapikey.com/v1",
  apiKey: process.env.ONEAIAPIKEY_API_KEY,
});

const stream = await client.chat.completions.create({
  model: "muse-spark-open",
  messages: [{ role: "user", content: "Write a TypeScript endpoint that returns paginated orders for the signed-in user." }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
}

More: quickstart · API reference · try it in Chat first

Benchmarks reported by the vendor

Scores for Muse Spark Open as its maker published them, with the setting they state. We have not re-run them. Versions and settings differ between vendors, so compare only identical benchmark names; compare models side by side.

Tokens per day, by model

Last 7 days, UTC days

Hover, tap or drag across any day, or focus the chart and use the arrow keys, to see how every model did that day.

Real tokens served through this gateway, updated every 5 minutes. Ranked by usage, not quality.

Full ranking insights

Muse line-up

Muse Spark Open or Muse Wonder or Mistral Large 4?

Of the 2 Muse models on this site, Muse Wonder is the step up from Muse Spark Open, and nothing in the family sits below it. Mistral Large 4 from Mistral AI is the closest-priced balanced-tier model from another vendor. Each comparison lists the published rates and limits, then what favours each side.

Muse Spark OpenvsMuse Wonder

Per 1M tokensMuse Spark OpenMuse Wonder
Input$0.87569% less$2.80
Output$2.9879% less$14
Cached input$0.1525% less$0.20
Context1.05M0% smaller1.05M

Pick Muse Spark Open when

  • cost per request decides it: Muse Spark Open charges input 69% less and output 79% less than Muse Wonder.
  • the work is coding, agent workflows and tool calling, which this catalogue lists for Muse Spark Open and not for Muse Wonder.

Pick Muse Wonder instead when

  • Muse Spark Open falls short and you want the flagship model of the pair.
  • inputs outgrow the 1.05M-token window of Muse Spark Open; Muse Wonder accepts 1.05M.
  • your workload looks like agent tasks, multimodal input and long documents, labels the catalogue gives Muse Wonder only.

Muse Spark OpenvsMistral Large 4

Per 1M tokensMuse Spark OpenMistral Large 4
Input$0.87584% more$0.476
Output$2.982.1× as much$1.43
Cached input$0.157% more$0.14
Context1.05M2× larger524.3k

Pick Muse Spark Open when

  • prompts run past 524.3k tokens: Muse Spark Open reads up to 1.05M.
  • the work is agent workflows and tool calling, which this catalogue lists for Muse Spark Open and not for Mistral Large 4.

Pick Mistral Large 4 instead when

  • the job runs well enough on the lower rates: input at 46% less and output at 52% less.
  • Muse Spark Open does not list Prompt caching and your code depends on it.
  • your workload looks like tool use and structured output, labels the catalogue gives Mistral Large 4 only.

Use cases

What Muse Spark Open is listed for, and what it costs

The catalogue lists Muse Spark Open for coding, agent workflows and tool calling. Each row prices one typical request at the published rates, with the arithmetic shown. The market-rate figure shows the same request at the lowest public price elsewhere. Token counts are examples; your prompts will differ.

Model your own traffic

  1. Coding

    Implementing a function or endpoint: 8,000 tokens of context in, 2,000 tokens of code out.

    Capable across languages for implementation, debugging and explaining code.

    “Write a TypeScript endpoint that returns paginated orders for the signed-in user.”

    Same request on Muse Wonder $0.0504 and Mistral Large 4 $0.006664.

    $0.01295

    per task, about 772 per $10

    $0.0185 at the market rate

    0.008 × $0.875 + 0.002 × $2.98

  2. Agent workflows

    One step of a tool-using agent: 20,000 tokens of instructions, tool results and history, 15,000 of them cached, and 1,000 tokens of tool calls and notes out.

    “Check the order status for each customer in this list and draft a reply to the late ones.”

    Same request on Muse Wonder $0.031 and Mistral Large 4 $0.005908.

    $0.0096

    per agent step, about 1,041 per $10

    $0.01275 at the market rate

    0.005 × $0.875 + 0.015 × $0.15 + 0.001 × $2.98

  3. Tool calling

    A request carrying 15 tool definitions: 5,000 tokens in, a 300-token tool call out.

    Define tools with precise JSON schemas; it picks and fills them reliably.

    “Find open invoices older than 30 days and draft a reminder to each customer.”

    Same request on Muse Wonder $0.0182 and Mistral Large 4 $0.002808.

    $0.005268

    per tool call, about 1,898 per $10

    $0.007526 at the market rate

    0.005 × $0.875 + 0.0003 × $2.98

Overview

When Muse Spark Open fits

Muse Spark Open is a balanced model from Meta, available here under the exact model id “muse-spark-open”. It is the entry point of the Muse line-up here, below Muse Wonder. Meta has not published a release date we can cite; it has been available here since 18 Oct 2026.

Meta documents tools, vision, reasoning and structured output for it, and it reads images as well as text.

It costs $0.875 per million input tokens and $2.98 per million output tokens, so output is 3.4× the input rate. That is 30% below the market rate of $1.25 input and $4.25 output per million tokens, the lowest public price we found at other API gateways. On the use cases below, output makes up 46% of the bill for coding, 31% for agent workflows and 17% for tool calling, so trimming the context you send saves more than capping the answer. With a 3:1 input-to-output mix it ranks 16th lowest-cost of the 27 models on this site.

Cached input bills at $0.15 per million tokens, 83% below the standard input rate. In the agent workflows example, 15k of the 20k input tokens come from cache, which brings one agent step from $0.020476 down to $0.0096.

It is served through the OpenAI-compatible format at https://api.oneaiapikey.com/v1, so any SDK or tool that lets you set a custom base URL, for example Codex CLI, Cursor and Cline, can use it. It is not offered on the Anthropic Messages endpoint; pick an Anthropic-format model for Claude Code.

Similar models

Same tier, other vendors

Same key, same balance. Switching is a change to the model field.

Compare side by sideAll models

Use Muse Spark Open with any balance.

Muse Spark Open is a paid model: any balance above $0 pays for it, the $10.00 sign-up credit included. When it runs out, add $1 or more to your wallet. Every request is billed at the rates above, with a receipt in your usage log.

Create an accountSee pricing

Muse Spark Open is developed by Meta. Model names and trademarks belong to their owners. One AI API Key is an independent API gateway and is not affiliated with, endorsed by or sponsored by Meta.