What is Muse Spark 1.3?
Muse Spark 1.3 is Meta’s proprietary hosted model model for long-horizon agentic workflows, coding, multimodal analysis and tool-assisted professional work. It was released September 3, 2026. This guide separates documented specifications from vendor benchmark claims and gives teams a practical way to decide whether the model belongs in a real evaluation.
QUICK VERDICT: Muse Spark 1.3 is Meta’s frontier model and ranks third in our verified leaderboard. Its price and million-token context make it a candidate, but Meta’s claims still require workload-specific testing.
Field | Verified value |
|---|---|
Provider | Meta |
Release date | September 3, 2026 |
Availability | Available through Muse Code and Meta Model API; max reasoning was still pending additional safety testing at launch |
License | Proprietary hosted model |
Context window | 1 million tokens |
Maximum output | Endpoint-specific within the documented context budget; verify the active Meta Model API limit |
Modalities | Text, image, video and document input; text and tool-directed output |
API pricing | Standard endpoint: $1.25 input, $0.15 cached input and $4.25 output per million tokens; contributor endpoint: $0.10 input, $0.002 cached input and $0.20 output |
Access | Muse Code and Meta Model API |
Best fit | long-horizon agentic workflows, coding, multimodal analysis and tool-assisted professional work |
Last verified | September 3, 2026 |
What changed with Muse Spark 1.3
Muse Spark 1.3 should not be evaluated as a name change. The important differences are the model’s reasoning controls, tool behavior, context policy, access route and economic profile. Those details determine whether a benchmark result transfers to production.
Area | What changed or matters |
|---|---|
Agentic workflows | Better context tracking, plan repair, clarification and handling of multiple workflows in one long thread |
Coding efficiency | Meta reports about 20% fewer tool calls and 25% fewer tokens than Muse Spark 1.2 |
Instruction following | Improved preservation of detailed requirements across multi-step work |
Safety | Stronger prompt-injection resistance and better recognition of irreversible actions |
Reasoning modes | Existing modes launched immediately; max reasoning was scheduled after additional safety testing |
The correct comparison baseline is Muse Spark 1.2. Teams planning a new deployment should also include Muse Spark 1.3 where access allows. Testing only the newest model or only the incumbent hides the migration cost, output-style changes and tool-use regressions that often matter more than a small benchmark gap.
Muse Spark 1.3 capabilities
The clearest fit is long-horizon agentic workflows, coding, multimodal analysis and tool-assisted professional work. That does not mean every task in those categories should use the model. A production system combines the base model with prompts, retrieval, tools, permissions, memory, retry logic and human review. The model card describes only one layer of that system.
Capability | Practical implication |
|---|---|
Long context | 1 million tokens. Test retrieval accuracy at realistic lengths rather than assuming every token receives equal attention. |
Output capacity | Endpoint-specific within the documented context budget; verify the active Meta Model API limit. Long output is useful only when the verification process can keep up. |
Modalities | Text, image, video and document input; text and tool-directed output. Confirm format-specific accuracy with your own files. |
Agent use | Use explicit tool schemas, narrow permissions, approval gates and recoverable operations. |
Reasoning | Record the exact effort level because cost, latency and quality can change materially. |
For coding, evaluate repository navigation, test creation, regression rate, review burden and recovery after a failed tool call. For research, measure citation correctness, source coverage and whether the model distinguishes evidence from inference. For computer or browser use, record every unsafe click, wrong field, lost state and unapproved action.
Benchmarks and evidence
The evidence below comes from Meta or named launch partners. It is useful for identifying intended strengths, but it is not equivalent to an independent head-to-head test. Prompts, tools, reasoning budgets, sampling, infrastructure and scoring rules can differ.
Evidence | Reported result | How to interpret it |
|---|---|---|
Current Overall ranking | 81.6 and number three in The AI Leaderboard’s September 3 top eight | LiveBench-derived comparative score across seven evaluation categories, not a production success rate |
Coding efficiency | About 20% fewer tool calls and 25% fewer tokens than Muse Spark 1.2 in Meta engineer comparisons | Vendor-reported internal comparison |
Long-horizon behavior | Meta reports better plan correction, context tracking, clarification and user collaboration | Provider capability claim that depends on the agent harness |
Evaluation coverage | Meta documents software engineering, terminal and multimodal evaluation methods | First-party methodology; inspect prompts, tools and scoring before comparison |
Treat benchmark results as a shortlist signal. Before purchasing or migrating, reproduce representative work under one frozen protocol. Keep model snapshots, reasoning effort, tool access and token budgets constant. Report repeated-run pass rates and total cost per accepted result, not a single best attempt.
Pricing, access and deployment
Muse Code and Meta Model API. The current pricing reference is Standard endpoint: $1.25 input, $0.15 cached input and $4.25 output per million tokens; contributor endpoint: $0.10 input, $0.002 cached input and $0.20 output. Provider pricing changes frequently and can include cache rates, batch discounts, regional premiums, long-context multipliers, priority processing and tool-call fees. Recheck the official pricing page before budgeting.
Cost driver | What to measure |
|---|---|
Input tokens | Prompt, retrieved context, tool results and repeated history |
Output tokens | Visible answer plus any billable reasoning or generated artifacts |
Caching | Eligible repeated prefixes, cache-read price and expiration policy |
Tools | Search, computer use, code execution and third-party API fees |
Retries | Failed runs, verifier loops and human rework |
Success-adjusted cost | Total spend divided by deliverables that pass review |
A cheaper token price can lose to a more expensive model if it takes more steps, retries more often or produces work that needs heavy correction. Conversely, a frontier model can be wasteful when a smaller model already passes the task rubric. Route by measured task difficulty rather than brand prestige.
Muse Spark 1.3 limitations
- Muse Spark 1.3 is proprietary and available only through Meta-controlled access routes.
- Max reasoning was not available at launch and was pending additional safety testing.
- Meta’s efficiency and capability comparisons are vendor-reported and need independent reproduction.
- A one-million-token context window does not guarantee reliable retrieval or instruction retention across the full window.
- Tool use, browser access and long-running coding workflows require least privilege, approval gates and recoverable operations.
High-stakes medical, legal, financial, security and scientific work requires qualified review. Store the exact model identifier, prompt version, tools, source documents and approvals for each consequential run. Build a rollback path before granting write access to repositories, browsers, databases or cloud infrastructure.
Risk | Minimum control |
|---|---|
Hallucination | Require source checks or executable tests |
Prompt injection | Separate untrusted content from instructions and restrict tools |
Over-permission | Use least privilege and approval gates |
Silent model change | Pin snapshots where possible and run regression tests |
Data exposure | Review retention, regional processing and provider terms |
Runaway cost | Set token, time, tool-call and retry budgets |
How to evaluate Muse Spark 1.3
Create 20 to 50 tasks from real work. Freeze the tasks and rubric before testing. Include easy tasks, normal tasks, edge cases and adversarial inputs. Give each model equivalent tools and enough budget to finish, but cap time and retries. Repeat non-deterministic runs so one lucky result does not decide the winner.
Test area | Record |
|---|---|
Task completion | Pass or fail plus rubric score |
Reliability | Repeated-run success and variance |
Quality | Factuality, instruction adherence and usefulness |
Tool use | Wrong calls, retries, recovery and permission errors |
Efficiency | Wall time, tokens, cache use, tool fees and human review |
Safety | Unsafe actions, injection response and sensitive-data handling |
Migration | Prompt changes, integration work and output-style regressions |
Compare Muse Spark 1.3 with Muse Spark 1.2 and Muse Spark 1.3. Choose the least expensive configuration that meets the acceptance threshold with an adequate safety margin. Re-run the suite after any model snapshot, system prompt, retrieval or tool change.
Frequently asked questions
Is Muse Spark 1.3 available now?
Available through Muse Code and Meta Model API; max reasoning was still pending additional safety testing at launch. The documented access routes are Muse Code and Meta Model API. Availability can differ by region, plan and partner platform.
Is Muse Spark 1.3 open source?
No. Muse Spark 1.3 is proprietary. Access is controlled by Meta and supported distribution partners.
How much does Muse Spark 1.3 cost?
Standard endpoint: $1.25 input, $0.15 cached input and $4.25 output per million tokens; contributor endpoint: $0.10 input, $0.002 cached input and $0.20 output. Budget with measured end-to-end workloads because token rates alone omit retries, tools, caching, long-context multipliers and review time.
What is Muse Spark 1.3 best used for?
Its strongest documented fit is long-horizon agentic workflows, coding, multimodal analysis and tool-assisted professional work. Start with a supervised pilot and keep human sign-off for consequential work.
Should I migrate from Muse Spark 1.2?
Only after a side-by-side evaluation. Measure pass rate, total cost, latency, output style, tool reliability and the engineering work required to migrate. A newer model is not automatically the better operational choice.
Related model guides
Start with the broader Best AI Models directory.
Compare with the Muse Spark 1.1 guide.
Compare with the Meta models hub.
Official sources and update policy
Meta Muse Spark 1.3 announcement.
Muse Spark 1.3 evaluation methodology.
Meta Model API access and pricing.
Checked September 3, 2026. We update this guide when the provider changes the model specification, pricing, access, licensing or safety documentation. Vendor benchmarks are attributed and are not presented as independent testing. No vendor payment or affiliate relationship determined inclusion.