Meta releases Muse Ember, a new lightweight model for agent execution
At 10 cents per million input tokens, Ember is significantly cheaper than frontier competitors like Haiku and Luna.
via TechCrunchReported by 2 outlets
Meta launches Muse Ember to tackle high-volume AI agent tasks at lower costs
The story so far
AI summary of the reporting below · Updated · How we work
Meta Superintelligence Lab launched Muse Ember on Monday, marketing it as its lightest and most cost-effective artificial intelligence model to date. Priced at 10 cents per million input tokens, the new release is aimed at complex reasoning execution and repetitive agent workloads such as tool calls, coding subtasks, routing, extraction, and long document parsing. [2]
The model is built to handle multiple formats natively, allowing agents to process text, audio, video, charts, screenshots, and scanned documents without turning to a secondary model. Meta designed Ember to work in tandem with heavier frontier systems like Muse Wonder, which can set operational rules while Ember executes them across large case volumes. [2]
Meta is pitching Ember directly against budget offerings from competitors, using Artificial Analysis benchmark scores to present it as both smarter and cheaper than rival models like OpenAI's GPT-6 Luna and Anthropic's Haiku. The release arrives amid intensifying industry rivalry, as competitors also challenge OpenAI with new frontier releases like Anthropic's Claude Fable 5.2. [1][2]
Read the original reporting
2 outlets
Read at TechCrunch
OpenAI Squeezed At The Frontier-End By Anthropic’s Claude Fable 5.2, While Meta’s Muse Ember Gives Its Bargain-Class GPT-6 Luna A Bruise
Today is not a good day for OpenAI, especially as it is getting hammered from two sides, with Anthropic's Claude Fable 5.2 leaving GPT-6 Astra in the proverbial dust, and Meta's Muse Ember squeezing GPT-6 Luna hard in the bargain category of AI models.
via Wccftech