Muse Spark 1.3
Last verified September 14, 2026 · Meta pricing ↗$1.25/1M input · $4.25/1M output · 1.0M context · Meta
Muse Spark 1.3 costs $1.25 per 1M input tokens and $4.25 per 1M output tokens from Meta, with a 1.0M-token context window. A typical chat reply (1,500 tokens in, 400 out) costs $0.0036, or $35.75 per 10,000 requests. On that workload it is cheaper than 70 of the 167 other buyable models tracked here.
Count Tokens
—
Tokens
—
Input Cost
—
Output Cost
Estimate Monthly Cost
Monthly Cost Estimator
Quick:
<$0.0001/mo
Pick a preset above or enter custom usage
What Muse Spark 1.3 Costs Per Request
At $1.25 per 1M input and $4.25 per 1M output tokens, before caching or batch discounts.
| Workload | Input / output tokens | Per request | Per 10,000 requests |
|---|---|---|---|
| Chat reply | 1,500 / 400 | $0.0036 | $35.75 |
| RAG answer over retrieved docs | 8,000 / 600 | $0.013 | $125.50 |
| Coding-agent step | 30,000 / 2,000 | $0.046 | $460.00 |
Muse Spark 1.3 vs Similarly Priced Models
Nearest buyable models by price, with the RAG answer above (8,000 in / 600 out) priced on each.
| Model | Input /1M | Output /1M | Context | RAG answer |
|---|---|---|---|---|
| Muse Spark 1.3 Meta | $1.25 | $4.25 | 1.0M | $0.013 |
| Muse Spark 1.1 Meta | $1.25 | $4.25 | 1.0M | $0.013 |
| o4 Mini OpenAI | $1.10 | $4.40 | 200K | $0.011 |
| Mistral Large 4 Mistral | $1.36 | $4.18 | 1.0M | $0.013 |
Alternatives to Muse Spark 1.3
Pricing Details
MetaLaunched September 2, 2026 (Axios dateline 2026-09-02; OpenRouter listing created 2026-09-02 19:45 UTC; TechRepublic prints September 4) into Muse Code and the Meta Model API, with Meta AI, Instagram and Facebook to follow. Model ID muse-spark-1.3. SAME RATE CARD AS MUSE SPARK 1.2 AND 1.1 TO THE CENT: $1.25 input, $4.25 output, $0.15 cached input (88% off), web search grounding $2.50 per 1,000 queries, 3,000 RPM / 4M TPM per team, 600 background submissions per minute per team. The difference from the 1.2 entry is provenance, not price: Meta now publishes these figures itself at ai.developer.meta.com/docs/pricing-rate-limits (verified 2026-09-14), which lists muse-spark-1.3, 1.2 and 1.1 side by side at one standard rate with no deprecation of either older model. Standard tier data clause on that page: 'your prompts and completions are not used to train Meta models'; the model page says 'Not used to improve our products'. OpenRouter (meta/muse-spark-1.3) and LiteLLM (meta/muse-spark-1.3) carry the identical $1.25/$4.25/$0.15 and 1,048,576 context. Max output is not on Meta's page: LiteLLM says 131,072, OpenRouter's top provider 943,718; the 16,384 stored here is carried over from the 1.1/1.2 entries. Meta claims ~20% fewer tool calls and ~25% fewer tokens than 1.2 on internal coding workflows, but Artificial Analysis (via TechRepublic) measured its per-task evaluation cost RISING from $0.40 on 1.2 to $0.55 on 1.3 from heavier input consumption. BENCHMARKS ARE FOR THE MAX REASONING SETTING, WHICH IS NOT SHIPPING: Meta's table compares 'Muse Spark 1.3 (max)' against 'Muse Spark 1.2 (xhigh)', GPT-5.6 Sol (max) and Opus 5 (max), while the max tier is held back for further safety testing; the API serves xhigh. Meta-published, 1.3 max vs 1.2 xhigh vs Sol vs Opus 5: Terminal-Bench 2.1 88.8 / 82.9 / 88.8 / 86.7; DeepSWE v1.1 75.4 / 55.0 / 73.0 / 74.0; SWEAtlas CodeBase QnA 59.4 / 46.2 / 53.5 / 52.7; MRCR 256K-512K 98.5 / 66.3 / 91.5 / -; MRCR 512K-1M 98.1 / 55.5 / 73.8 / -; JobBench 64.9 / 61.6 / 45.4 / 65.7; OSWorld 2.0 66.9 partial, 32.0 binary / 47.6, 17.9 / 62.7, 27.3 / 68.3, 31.4; DeepSearchQA 90.3 / 85.9 / 93.1 / 90.4; Agentic IF Index (internal) 57.8 / 46.2 / 60.5 / 59.1; AutomationBench 49.6 / 38.2 / 46.7 / 50.3; GDPVal-AA v2 Elo 1754 / 1615 / 1710 / 1824. Meta's own table now prints 1.2 DeepSWE at 55.0, against the 59.3 Meta gave at the 1.2 launch. Artificial Analysis Intelligence Index 61 for the shipping xhigh, 62 for max in limited preview (via DataCamp). Estimated tokens (different tokenizer).
Input / 1M tokens
$1.25
Output / 1M tokens
$4.25
Context Window
1.0M
Max Output
16K
Price History
No price change recorded since 2026-09-14, when this site started tracking this model; it launched on 2026-09-02.
Frequently Asked Questions
Common questions about Muse Spark 1.3 pricing and usage
How much does Muse Spark 1.3 cost?
Muse Spark 1.3 costs $1.25 per 1 million input tokens and $4.25 per 1 million output tokens. This pricing is set by Meta.
What tokenizer does Muse Spark 1.3 use?
Muse Spark 1.3 uses the cl100k_base tokenizer. This tokenizer has a vocabulary of ~100,000 tokens and is widely used across many models.
What is Muse Spark 1.3's context window?
Muse Spark 1.3 supports a context window of 1.0M tokens (1,048,576 tokens). The maximum output length is 16K tokens.
How does Muse Spark 1.3 compare to other models?
On an 8,000-token-in, 600-token-out RAG answer, Muse Spark 1.3 costs $0.013 per request. The nearest-priced alternatives cost: Muse Spark 1.1 (Meta) $0.013, o4 Mini (OpenAI) $0.011, Mistral Large 4 (Mistral) $0.013. The table above lists their full rates and context windows.
How do I count tokens for Muse Spark 1.3?
Use the token counter at the top of this page or visit the homepage calculator. Paste your text and it will show the exact token count and cost for Muse Spark 1.3 using the cl100k_base tokenizer.
Read More About Muse Spark 1.3
Muse Spark Open pricing: five hosts, one price, and it is closed Muse Spark 1.3's card for a model that scores six points lower
October 17, 2026 · 7 min read
Meta priced Muse Wonder at Claude Opus 5.5's $4 and $20 to the cent. Opus on high effort outscores it for 40% of the cost per task, and the Wonder that beats Opus at coding isn't sold by the token.
October 13, 2026 · 6 min read