Kavya Kopparapu

research scientist @meta superintelligence • former research @googledeepmind / cs + biology @harvard / founder @girlscomputing

[Mark Zuckerberg @finkd]

(1) Today we're releasing Muse Spark 1.1 -- a strong agentic and coding model at a very low price. It's available through our new Meta Model API and in Meta AI.

[Alexandr Wang @alexandr_wang]

1/ muse spark 1.1 is an industry-competitive agentic and coding model. across many agentic evals it rivals gpt-5.5 and opus-4.8.

available now through the new meta model api and in meta ai. 🧵

[Shengjia Zhao @shengjia_zhao]

Today we are launching Muse Spark 1.1, an upgrade to muse spark 1 that greatly improves agentic, coding, multimodal, and computer use capabilities. We're also launching the Meta Model API in public preview.

https://t.co/Th9DzacJaW

[Shengjia Zhao @shengjia_zhao]

Excited to share what we’ve been building at Meta Superintelligence Labs! We just released Muse Spark, our first AI model. It's a natively multimodal reasoning model and the first step on our path to personal superintelligence. We've overhauled our entire stack to support scaling, and this is just the beginning.

https://t.co/KNVjgMcch1

[Vals AI @ValsAI]

Meta just released a frontier model, Muse Spark- it takes the #3 spot on our Vals Index.

[Artificial Analysis @ArtificialAnlys]

Meta is back! Muse Spark scores 52 on the Artificial Analysis Intelligence Index, behind only Gemini 3.1 Pro, GPT-5.4, and Claude Opus 4.6. Muse Spark is the first new release since Llama 4 in April 2025 and also Meta's first release that is not open weights

Muse Spark is a new model from @Meta evaluated on Artificial Analysis. We were given early access by Meta to independently benchmark the model. It is the first frontier-class model from Meta since Llama 4 Maverick was released in April 2025, and notably the first @AIatMeta model that is not being released as open weights. The release follows Meta's reorganization of its AI efforts under Meta Superintelligence Labs, and signals that Meta is re-entering the frontier race after roughly a year of relative quiet.

For context, Llama 4 Maverick and Scout scored 18 and 13 respectively on the Artificial Analysis Intelligence Index as non-reasoning models at the time of their release, while Muse Spark scores 52. Muse Spark essentially closes the gap between to the frontier in a single release.

The model is not open source and is not yet accessible via an API but Meta has shared they expect this to come soon. Meta is also integrating Muse Spark into their first party products including their Meta AI chat product, Facebook, Instagram and Threads.

Key takeaways from our benchmarks:
➤ Muse Spark scores 52 on the Artificial Analysis Intelligence Index, placing it within the top 5 models we have benchmarked. It sits ahead of Claude Sonnet 4.6, GLM-5.1, MiniMax-M2.7, Grok 4.20 and behind Gemini 3.1 Pro Preview, GPT-5.4 and Claude Opus 4.6

➤ Muse Spark is notably token efficient for its intelligence level. It used 58M output tokens to run the Intelligence Index, comparable to Gemini 3.1 Pro Preview (57M) and notably lower than Claude Opus 4.6 (Adaptive Reasoning, max effort, 157M), GPT-5.4 (xhigh, 120M) and GLM-5 (110M)

➤ Muse Spark is the second-most capable vision model we have benchmarked. It scores 80.5% on MMMU-Pro, behind only Gemini 3.1 Pro Preview (82.4%)

➤ Muse Spark performs strongly on reasoning and instruction-following evaluations. It scores 39.9% on HLE, trailing only Gemini 3.1 Pro Preview (44.7%) and GPT-5.4 (xhigh, 41.6%). The model also achieved 5th highest in CritPT with a score of 11%, an eval that is focused on difficult physics research questions. This is substantially above above Gemini 3 Flash (9%) and Claude 4.6 Sonnet (3%)

➤ Agentic performance does not stand out. On GDPval-AA, our evalaution focused on real world work tasks, Muse Spark scores 1427, behind both Claude Sonnet 4.6 at 1648 and GPT-5.4 at 1676, but ahead of Gemini 3.1 Pro Preview at 1320. On On TerminalBench Hard, Muse Spark trails Claude Sonnet 4.6, GPT-5.4, and Gemini 3.1 Pro. Muse Spark joins others in achieving a high τ²-Bench Telecom score of 92%

Key model details:
➤ Modalities: Multimodal including text and vision input, text output
➤ License: Proprietary, Meta's first frontier model not released as open weights
➤ Availability: No public API at the time of publishing. Meta expects to provide API access soon. Meta has started integration into their first party AI offering Meta AI and inside Facebook, Instagram, and Threads

[Alexandr Wang @alexandr_wang]

1/ today we're releasing muse spark, the first model from MSL. nine months ago we rebuilt our ai stack from scratch. new infrastructure, new architecture, new data pipelines. muse spark is the result of that work, and now it powers meta ai. 🧵

[Kavya Kopparapu @kavyakvk]

@FujifilmX_US Did they drop yet? I can’t tell if the out of stock notification is because they’re all gone or because they haven’t released the units yet

[Kavya Kopparapu @kavyakvk]

@FujifilmX_US Thanks!! 🤞

[elvis @omarsar0]

Self-Improvement for Multi-Step Reasoning LLM Agent

Proposes a ReAct-style agent with self-critique for improving on the task of long-form question answering.

It shows that the agent can be improved through ReST-style (reinforced self-training) iterative fine-tuning on its reasoning traces. Specifically, it uses growing-batch RL with AI feedback for continuous self-improvement and self-distillation.

Like a few other recent papers, it focuses on minimizing human involvement (i.e., doesn't rely on human-labeled training data). It generates synthetic data with self-improvement from AI feedback which can then be used to distill the agent into smaller models (1/2 orders magnitude) with comparable performance as the pre-trained agent.

Great paper with interesting ideas of how future LLM agents could be improved and how to deal with challenges like obtaining multi-step human-labeled data at scale.

https://t.co/OAHXDMyOXR

[Kavya Kopparapu @kavyakvk]

@kaiwang_gua @GeorgiaTech @MilindTambe_AI Congrats Kai!!

[Mihir Patel @mvpatel2000]

We're back 😈. Stable diffusion in $50k, and all the code is open source https://t.co/KQkdozNzd5. Plus, this time I finally convinced @jefrankle to give me enough GPUs to train it, so we have the model to prove it.

https://t.co/EQ8fQk3Bzp

[Mihir Patel @mvpatel2000]

Imagine training GPT 3 from scratch at only $450k with our infra. Now imagine how much it actually costs with our speedups... 🤫😏

[HP @HP]

The kids are alright: During a period in history where it is easy to be pessimistic about the planet’s future, meet 16 #GenZ activists, organizers, and advocates who are much-needed sources of hope ✊ https://t.co/a0kklaPd9f @gendercool @EllyanneCGithae
@fsemmha @KavyaKopparapu

[Girl Up @GirlUp]

The Global Leadership Summit presented by @ProcterGamble is going to be 🔥🔥🔥 #GirlsLead21

Welcome to the lineup @SonyaDevilleWWE @LinaAbiRafeh @rosaliefishx @tabataamaralsp @KavyaKopparapu!

[Kavya Kopparapu @kavyakvk]

Applications for the @girlscomputing scholars program are open! If you have a passion for teaching low-income students about emerging tech (and getting a scholarship along the way) apply at https://t.co/jlkOOiNRWu

[Howard Lerman @howard]

TJHSST Students and Alums, let's connect this Thursday night 9:00pm EST!

https://t.co/JTYVnglrMA

[Entrepreneur Middle East @EntMagazineME]

.@KavyaKopparapu: “Healthcare is not a one-sided problem. There’s an underlying real-world impact. The inventor, whoever is driving forward the change, needs to understand all aspects. Technology, and innovation in particular, always reflects the innovator." #UAE_innovates

[Kavya Kopparapu @kavyakvk]

🚀

[Entrepreneur Middle East @EntMagazineME]

When talking about moonshot goals, @KavyaKopparapu says it’s important to look for value-driven goals. “How does your goal impact more people?” #UAE_innovates #innovate4health