Meta model product
2 is Meta Superintelligence Labs' proprietary multimodal reasoning model for long-horizon coding and whole-repository agentic work.
Updated Aug 12, 2026. Default version: Muse Spark 1.2
Technical details for the model's default version.
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Benchmark scores for Muse Spark 1.2.
| Meta Internal Coding Bench | 70.6% | 01 | 1 | 100.0% | C | |
| DeepSWE 1.1 | 59.3% | 08 | 20 | 63.2% | C | |
| Terminal-Bench 2.1 | 82.9% | 08 | 19 | 61.1% | C |
Preference and agent-evaluation results for the default version.
| text style control | overall | 04 | 1498.6 | 3,278 | N/A | |
| text factuality | overall | 07 | 1485.2 | 3,278 | N/A | |
| text | overall | 08 | 1488.2 | 3,278 | N/A | |
| vision style control | overall | 10 | 1290.0 | 1,282 | N/A | |
| vision | overall | 13 | 1302.3 | 1,282 | N/A | |
| webdev | overall | 19 | 1535.4 | 2,053 | N/A |
Provider-specific output speed and catalog latency for Muse Spark 1.2. Runtime does not affect the capability score.
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Official vendor API pricing appears first, followed by individual provider offers.
| NanoGPT | meta/muse-spark-1.2 | global | $1.25 | $4.25 | 1M | |
| LLM Gateway | muse-spark-1.2 | global | $1.25 | $4.25 | 1M | |
| Meta | muse-spark-1.2 | global | $1.25 | $4.25 | 1M | |
| Kilo Gateway | meta/muse-spark-1.2 | global | $1.25 | $4.25 | 1M | |
| Merge Gateway | meta/muse-spark-1.2 | global | $1.25 | $4.25 | 1M | |
| Vercel AI Gateway | meta/muse-spark-1.2 | global | $1.25 | $4.25 | 1M | |
| OpenRouter | meta/muse-spark-1.2 | global | $1.25 | $4.25 | 1M |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Available versions of this model. The score column identifies the version used in the overall ranking.
| Muse Spark 1.2 | 82.3 | N/A | 1M | 131.1K | No | Proprietary |
Key information about Muse Spark 1.2 and its available data.
2 is Meta Superintelligence Labs' proprietary multimodal reasoning model for long-horizon coding and whole-repository agentic work. It supports a 1,048,576-token context window and is available through the Meta Model API in standard and contributor tiers.
The contributor endpoint offers discounted inference in exchange for allowing prompts and completions to be used to improve Meta's products; ZeroEval routes to that endpoint while displaying the standard tier's data-private pricing.
Data as of 2026-08-11.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
Common questions about Muse Spark 1.2.
Muse Spark 1.2's default version was released on Aug 5, 2026.
Muse Spark 1.2's official API price is $1.25 per million input tokens and $4.25 per million output tokens via Meta. The lowest tracked third-party offer starts at $1.25 input and $4.25 output via NanoGPT.
Muse Spark 1.2 was created by Meta.
The default version has a 1M token context window.
No. The default version is not marked as having publicly available weights.
7 provider offerings are linked to the default version.
Nearby ranked alternatives include Claude Opus 4.7, Claude Sonnet 5, GPT-5.5.