After lagging behind top labs in AI capabilities, Meta formed the new Meta Superintelligence Labs (MSL) group to develop frontier-level AI models. Following months of working behind the scenes, Meta announced the Muse Spark model in April, which delivered competitive results compared to Opus 4.6 Max, Gemini 3.1 Pro High, GPT 5.4 xhigh, and Grok 4.2 Reasoning.
Today, Meta"s MSL announced Muse Spark 1.1, a major upgrade over the original Muse Spark, featuring improvements across agentic tasks, computer use, coding, and multimodal understanding. Meta highlighted that Muse Spark 1.1 can work with native tools, MCP servers, and custom skills. It can also act as a main agent that creates a plan and delegates work to parallel subagents.
The new Muse Spark 1.1 supports a 1-million-token context window, a significant upgrade from Meta"s earlier models. Meta claims that this model can actively manage that context during long workflows. Meta is also positioning Muse Spark 1.1 as a stronger model for computer-use tasks.
For the first time, Meta is making its model available through an API for building AI applications. At launch, new Meta Model API accounts will get $20 in free credits. After that, the model is priced at $1.25 input/$4.25 output per million tokens, which is significantly lower than other frontier models from Anthropic and OpenAI.
As you can notice from the benchmark comparison table below, Muse Spark 1.1 now holds state-of-the-art scores on MCP Atlas, JobBench, Humanity"s Last Exam, and FinanceBench. On other key benchmarks, it rivals much costlier models, including OpenAI"s GPT-5.5 and Anthropic"s Opus-4.8.
On the safety side, Meta says it evaluated Muse Spark 1.1 under its Advanced AI Scaling Framework. The company claims the model remains within safe margins across frontier risk areas such as chemical and biological risks, cybersecurity, and loss of control.
For general consumers, the new Muse Spark 1.1 model is available now in “Thinking” mode in the Meta AI app and on meta.ai.
Source: Meta