Meta has released Muse Wonder, saying it outperforms comparable Anthropic and OpenAI models on some technical benchmarks and can interpret leases and organize financial or health information. The model will be closed-source and priced in tiers, with basic usage starting at $0.2 per million tokens; its most capable version will be sold to businesses.
Meta 发布首批前沿模型 Muse Wonder 和 Muse Wonder Full
Meta 称,两款模型比此前的 Muse Spark 更大,在编码等复杂任务上表现更好;其测试显示,模型跑分与 Fable 5.1 和 GPT-6 Astra 相当。Wonder 可通过 Meta Model API 使用,定价为每百万输入 token 4 美元、输出 token 20 美元;Wonder Full 仅面向 Muse Code 订阅者,不通过 API 提供。
Event developments
Coverage of the same occurrence shares a node; subsequent developments have their own nodes.
Meta releases its first frontier models, Muse Wonder and Muse Wonder Full
Meta says the two models are larger than Muse Spark and perform better on complex tasks such as coding; in Meta’s tests, their benchmark scores were comparable to Fable 5.1 and GPT-6 Astra. Wonder is available through the Meta Model API at $4 per million input tokens and $20 per million output tokens. Wonder Full is limited to Muse Code subscribers and is not offered through the API.
View source reports · 2 sources · 3 reports
All reports · 3
By original publication time, with each report's bookmarks and feedback preserved.
Meta has released Muse Wonder and Muse Wonder Full. According to Meta's figures, both scored close to Anthropic's top models in reasoning, writing and coding tests. Muse Wonder is expected to power some Muse app features soon, while the more capable Full version is available only to Muse Code subscribers.
Meta says the two models are larger than Muse Spark and perform better on complex tasks such as coding; in Meta’s tests, their benchmark scores were comparable to Fable 5.1 and GPT-6 Astra. Wonder is available through the Meta Model API at $4 per million input tokens and $20 per million output tokens. Wonder Full is limited to Muse Code subscribers and is not offered through the API.