Muse Spark 1.3
PulseAugur coverage of Muse Spark 1.3 — every cluster mentioning Muse Spark 1.3 across labs, papers, and developer communities, ranked by signal.
- 2026-09-09 product_launch Meta released Muse Spark 1.3, a closed multimodal reasoning model with a 1 million token context window and dual pricing tiers. source
- 2026-09-04 product_launch Meta released its Muse Spark 1.3 model, designed for longer work that uses external tools. source
- 2026-09-02 product_launch Meta released its new AI model, Muse Spark 1.3. source
- 2026-09-02 product_launch Meta released Muse Spark 1.3, an AI model that matches or surpasses benchmarks set by GPT-5.6 Sol and Fable 5. source
- 2026-09-02 product_launch Meta released its new AI model, Muse Spark 1.3, which matches frontier model performance at a significantly lower cost. source
- 2026-09-02 product_launch Meta released Muse Spark 1.3, an AI model designed for agentic and coding tasks, offering competitive performance and a tiered pricing model. source
- 2026-09-02 product_launch Meta AI has launched Muse Spark 1.3, an AI model designed for agentic workflows and coding tasks. source
- 2026-09-02 product_launch Meta AI has released Muse Spark 1.3, an updated large language model. source
4 day(s) with sentiment data
Muse Spark 1.3 optimized for agentic, tool-use tasks, particularly in finance and coding.
Multiple clusters highlight Muse Spark 1.3's specialization in coding and agentic tasks, with specific mentions of banking and terminal operations. Its 1 million token context window and availability via Muse Code and Meta Model API suggest a focus on applications requiring interaction with external tools and handling large codebases or financial data.
Muse Spark 1.3's API-only nature will limit its adoption compared to more open models.
While Muse Spark 1.3 is noted for its performance in specific domains like banking and terminal operations, the cluster evidence explicitly states it remains a proprietary API-only model. This contrasts with other models that might offer more flexible deployment options, potentially hindering its widespread adoption by developers seeking greater control or integration flexibility.
Meta will release a frontier model in the next 30 days, potentially building on Muse Spark 1.3's capabilities.
One cluster mentions Meta teasing a new frontier model alongside the Muse Spark 1.3 release. Given the recent releases of competitor frontier models (Anthropic's Fable 5.1, OpenAI's GPT-6 Astra), it is plausible Meta will follow suit soon to remain competitive, possibly integrating advancements seen in Muse Spark 1.3.
-
Meta's Muse Wonder API sets team-wide rate limits of 3,000 RPM and 4M TPM
Meta's Muse Wonder API has established rate limits of 3,000 requests per minute (RPM) and 4,000,000 tokens per minute (TPM), applicable per model and per team rather than per API key. These limits are shared across all …
-
Meta's Muse Wonder pricing revealed: $4 in, $20 out per million tokens
Meta has released pricing details for its Muse Wonder model, with input tokens costing $4 per million and output tokens at $20 per million. Caching can reduce input costs to $0.20 per million tokens, and reasoning token…
-
MemStrata achieves 95% accuracy on long-context benchmarks with local Qwen reader
Researchers have developed MemStrata, a system that achieves high source-aware accuracy on long-context evaluation benchmarks. Using a local Qwen 3.8 27B Q4_K_M reader, MemStrata CL1 reached 95% accuracy on LongMemEval-…
-
Anthropic's Claude Opus 5.5 excels in video generation and benchmarks · 2 sources tracked
Anthropic has released Claude Opus 5.5, which has quickly garnered positive attention, particularly for its capabilities in generating explainer videos. The model demonstrates strong performance on various benchmarks, i…
-
Muse Spark 1.3 praised for coding, users await open-source release
The Muse Spark 1.3 model is receiving praise for its coding capabilities and speed. Users are expressing hope that the developers will release the model as open source, as was previously discussed. There is anticipation…
-
Meta's Muse Spark 1.3 tops Website Arena, setting new speed and cost benchmarks · 1 source tracked
Meta's Muse Spark 1.3 has achieved the top position on the Design Arena's Website Arena leaderboard, a benchmark that uses real-world user votes to evaluate AI models for website creation. This marks a significant advan…
-
Google and Meta release new coding-focused AI models
Google and Meta have both released new coding-focused AI models, Gemini 3.8 Flash and Muse Spark 1.3, respectively. These models are designed for high-volume coding and agentic tasks at a lower price point than flagship…
-
Meta's Muse Spark 1.3 launches with 1M context, dual pricing tiers
Meta has released Muse Spark 1.3, a closed multimodal reasoning model designed for long-running agentic tasks and coding. The model boasts a 1 million token context window and scores highly on benchmarks like DeepSWE v1…
-
User praises Meta's Muse Spark 1.3 for affordability and performance
A user expressed satisfaction with Meta's Muse Spark 1.3 Contributor, finding it to be an excellent and cost-effective tool. The user indicated a willingness to share their project data with Meta, specifically mentionin…
-
Cursor integrates Meta's Muse Spark 1.3 into AI IDE
Cursor has announced the integration of Meta's Muse Spark 1.3 model into its AI IDE. This integration allows users to leverage the capabilities of Muse Spark 1.3 directly within the Cursor development environment. The c…
-
Meta AI unveils Muse personal agent powered by Muse Spark 1.3
Meta AI has launched Muse, a new personal agent designed to assist users with tasks. This agent is powered by Muse Spark 1.3, a new underlying model. The company has provided a look into the development and capabilities…
-
SemiAnalysis flags Gemini 3.8 Flash and Muse Spark 1.3 as highly optimized models
SemiAnalysis has identified Gemini 3.8 Flash and Muse Spark 1.3 as highly optimized models based on their benchmarking. These models appear to have been specifically tuned to perform exceptionally well on evaluation metrics.
-
Public AI benchmarks lose value quickly as models optimize for them, says SemiAnalysis
SemiAnalysis argues that public AI benchmarks like TB 4.0 quickly become obsolete as models are trained to optimize for them, diminishing their usefulness for evaluating true generalization. They highlight Gemini 3.8 Fl…
-
AI Model Fatigue Sets In as Four Labs Release New Models in One Week
The AI industry is experiencing "model fatigue" as four major labs—Anthropic, Meta, Google, and OpenAI—released new models within a single week. This rapid release cycle, with a median gap of just 11 days between major …
-
AI model releases accelerate, creating a 're-evaluation tax' for builders
The AI model release cycle has accelerated dramatically, with four major models (Claude Fable 5.1, Gemini 3.8 Flash, Muse Spark 1.3, and GPT-6 Astra) being launched in just four days. This rapid pace creates a significant
-
Anthropic, Meta, and OpenAI release new AI models
A week saw the release of three significant AI models: Anthropic's Fable 5.1, Meta's Muse Spark 1.3, and OpenAI's GPT-6 Astra. These releases collectively offer insights into the future direction and capabilities of nex…
-
Artificial Analysis Index criticized for misrepresenting AI performance
The Artificial Analysis Index is being criticized for not accurately reflecting real-world AI performance. A user tested Muse Spark 1.3 and found it to be inferior to other models like OPUS and SOL, suggesting the index…
-
Meta releases Muse Spark 1.3 for tool-based AI tasks
Meta has released its Muse Spark 1.3 model, designed for extended tasks that leverage external tools. This new model is available through Muse Code and the Meta Model API, aiming to enhance productivity for developers a…
-
OpenAI's GPT-6 Astra excels at tasks, not general intelligence · 1 source tracked
OpenAI has released GPT-6 Astra, its latest frontier model, which shows significant improvements in specific task-oriented capabilities like long-horizon reasoning and discovering rules, rather than a general increase i…
-
Google releases rapid-fire Gemini Flash models amid flagship delay
Google has released four Gemini Flash models in rapid succession, with the latest, Gemini 3.8 Flash, excelling in coding and matching larger models' performance at a lower cost. This rapid release pace is highlighted by…