Meta Superintelligence Labs has released Muse Spark 1.1, a multimodal reasoning model designed for agentic tasks. Simultaneously, the company opened a public preview of the Meta Model API. This shift marks a structural change in how developers access Meta’s models. Previously, access came mainly through open weights. Muse Spark 1.1 is now closed, hosted, and metered per token. The practical question is where this fits into a stack you already run.
What is Muse Spark 1.1?
Meta describes it as a multimodal reasoning model built for agentic tasks. Reported gains over the first Muse Spark sit in tool use, computer use, coding, and multimodal understanding. The context window is 1,000,000 tokens. Meta Model API documentation lists 1,048,576.
Core Capabilities and Features
Because it is a reasoning model, it thinks before answering. This reasoning effort is adjustable per request. Inputs span text, images, video, and documents; output is text. The API also exposes structured output, parallel tool calling, a Files API, and prompt caching. Adding a web_search tool to a Responses API call returns cited answers.
Pricing and Regional Availability
Access splits two ways. Consumers get it free in ‘Thinking’ mode in the Meta AI app and on meta.ai. Developers pay $1.25 per million input tokens and $4.25 per million output tokens. New accounts get $20 in free credits. Initial launch posts describe the public preview as US-only, with no EU access yet.
Performance
With the spec in place, the numbers explain the positioning. To illustrate this, Meta published a launch table, and the table splits cleanly.
| Benchmark | Tests | Muse Spark 1.1 | Opus 4.8 (max) | GPT-5.5 (xhigh) | Gemini 3.1 Pro (high) |
|---|---|---|---|---|---|
| MCP Atlas | Scaled tool use | 88.1 | 82.2 | 75.3 | 78.2 |
| JobBench | Professional tool use | 54.7 | 48.4 | 38.3 | 15.9 |
| Humanity’s Last Exam | Reasoning with tools | 62.1 | 57.9 | 52.2 | 51.4 |
| OSWorld-Verified | Computer use | 80.8 | 83.4 | 78.7 | 76.2 |
| SWE-Bench Pro | Real-repo coding | 61.5 | 69.2 | 58.6 | 54.2 |
| DeepSWE 1.1 | Long-horizon coding | 53.3 | 59.0 | 67.0 | 12.0 |
| BabyVision | Visual reasoning | 76.3 | 81.2 | 83.6 | 51.5 |
Benchmark Analysis
Meta reported these figures with rivals shown in their strongest modes. Muse Spark 1.1 leads the tool-use and tool-augmented reasoning rows. It places third on coding and multimodal tasks. Consequently, this is an orchestration model, not a coding-accuracy leader. Meta also chose the benchmark set and ran the harness.
The Feature Worth Taking Seriously: Compaction and Delegation
Beyond the scores, orchestration behaviour explains the tool-use results. The model actively manages its million-token context window. It remembers actions, retrieves information from much earlier work, and compacts what it keeps.
Delegation is the second half. As a main agent, it gathers context, plans, and delegates execution across parallel subagents. As a subagent, it adheres to its job, understands available tools, and escalates back when needed. The research team also reports zero-shot generalization to new native tools, MCP servers, and custom skills.
Computer use follows the same logic. The model was trained to write scripts when automation is faster. It clicks when direct interaction is simpler. It generates batches of actions at each step.
Wiring It Into an Existing Stack
Because the Model API is OpenAI-compatible, migration is a base-URL change rather than a rewrite. The snippet below is Meta’s own first-call recipe.
# pip install openai
import os
from openai import OpenAI
# The OpenAI SDK does not auto-read MODEL_API_KEY, so pass it explicitly.
client = OpenAI(
base_url="https://api.meta.ai/v1",
api_key=os.environ["MODEL_API_KEY"],
)
response = client.chat.completions.create(
model="muse-spark-1.1",
messages=[{"role": "user", "content": "Hello, world!"}],
)
print(response.choices[0].message.content)Anthropic-format harnesses, such as Claude Code, point at the Messages API instead. Agent CLIs like OpenCode register a provider using three values: base URL, key, model ID.
Use Cases
In practice, the shipped demos map onto workloads engineering teams already recognise.
For instance:
Multimodal listing automation
- In the Facebook Marketplace demo, the model takes smartphone video, extracts useful photos, reasons about the product, then operates a browser to publish the listing.
Screenshot-driven debugging
- In an OpenCode demo, it builds a chat web app, takes automated screenshots, traces failures back to code, then validates fixes.
Adaptive planning
- In the dinner-party demo, new context arrives mid-order, and the model updates the plan unprompted.
Coding harnesses get first-class support: planning mode, goal conditioning, subagent delegation, context compaction. Meta team also reports a significant gain on Meta Internal Coding Bench.
Strengths and Weaknesses
Taken together, the trade is legible.
Strengths
- Leads Meta’s reported tool-use and tool-augmented reasoning evaluations
- Million-token context, compacted by the model across long sessions
- Zero-shot generalization to unfamiliar tools, MCP servers, custom skills
- OpenAI and Anthropic SDK compatibility makes an A/B test cheap
Weaknesses
- Third on SWE-Bench Pro, DeepSWE 1.1, and BabyVision in Meta’s table
- Closed weights, so no local deployment and no fine-tuning
- Every launch figure is vendor-reported, rivals shown at maximum settings
- Public preview reported as US-only; preview pricing can change
What it means
For developers, Muse Spark 1.1 is the first paid API for Meta’s own frontier model. It offers a drop-in replacement for existing OpenAI or Anthropic clients. The million-token window is actively compacted by the model itself, handling long sessions without manual context management. However, the lack of open weights means you cannot run it locally or fine-tune it. Pricing is set at $1.25 per million input tokens and $4.25 per million output tokens, with $20 in free credits for new accounts. The public preview remains US-only.
