Meta's Muse Spark 1.1 outperforms GLM-5.2 in coding and costs slightly less
Meta’s Muse Spark 1.1 improves coding performance by 8 points on the Artificial Analysis Intelligence Index and reduces hallucination rates by 35 percentage points while costing less than GLM-5.2.
- Muse Spark 1.1 scores 51 on the Artificial Analysis Intelligence Index, an 8-point improvement in three months.
- In coding benchmarks, it outperforms GLM-5.2 with a score of 71.3 while costing $0.26 per task.
- Hallucination rate drops from 73% to 38%, a 35 percentage point reduction.
- The update positions Meta’s model as a more cost-effective and reliable alternative for developers.
Meta has released Muse Spark 1.1, an updated version of its AI model that shows marked improvements in coding performance and cost efficiency. On the Artificial Analysis Intelligence Index, the model scored 51, an eight-point increase over its previous iteration in just three months. In coding-specific benchmarks, Muse Spark 1.1 achieved a score of 71.3, surpassing GLM-5.2, a rival model from Zhipu AI.
Beyond performance gains, the update also addresses a critical issue in AI reliability. The hallucination rate for Muse Spark 1.1 dropped dramatically from 73% to 38%, a 35 percentage point reduction that could significantly improve trust in automated coding assistance. Additionally, the model’s cost per task is slightly lower at $0.26, making it a more economical choice for developers and businesses.
The improvements come at a time when AI models are increasingly being evaluated not just on raw performance but also on cost and reliability. Muse Spark 1.1’s advancements suggest Meta is making strides in balancing these factors, which could influence adoption rates in competitive markets like software development and enterprise automation.
Lower cost and reduced hallucinations make Muse Spark 1.1 a practical choice for coding tasks.
Improved reliability and cost efficiency could drive adoption in enterprise automation and software development.
AI models are being judged on both performance and trustworthiness, and this update reflects that shift.
- Artificial Analysis Intelligence Index
- A benchmarking system that evaluates AI models across multiple dimensions, including coding performance and cost efficiency.
- Hallucination rate
- The percentage of incorrect or fabricated outputs generated by an AI model, a key metric for reliability.
OpenAI releases GPT-5.6 and ChatGPT Work tool - Axios
LLMOpenAI launches its new family of models with GPT-5.6
LLMOpenAI Releases GPT-5.6 (Sol, Terra, Luna): A Three-Tier Model Family With Programmatic Tool Calling in the Responses API
LLMOpenAI's GPT-5.6 and ChatGPT Work aim to beat Anthropic on price, speed, and productivity
LLMGPT-5.6 Sol nearly matches Fable 5 on aggregated benchmarks at one-third the cost
AI ResearchChina's Orca world model matches specialized robotics systems without ever seeing a single action label
China’s Beijing Academy of Artificial Intelligence unveiled Orca, a world model trained on 125,000 hours of video that predicts abstract states instead of tokens or pixels. It matches specialized robotics systems on five benchmark tasks without using any action labels.
AI moved in next door. For this Memphis community, life got more complicated. - The Christian Science Monitor
A Memphis community is grappling with the arrival of AI, leading to increased complexity in daily life. The Christian Science Monitor reports on the challenges faced by residents.
Macroscope | Don’t expect the rising tide of AI to lift all boats - South China Morning Post
A recent article by the South China Morning Post warns that the growing use of AI may not have a positive impact on all industries.
BusinessOpenAI admits it "didn't get everything quite right" with ChatGPT Work launch and scrambles to fix UX and costs
OpenAI has admitted to multiple issues with its new ChatGPT Work and GPT-5.6 Sol releases, including excessive compute usage, confusing desktop transitions, and unauthorized data deletions.
RoboticsAnt Group’s Robbyant Unveils LingBot-VA 2.0: A Causal Video-Action Model Built Natively for Physical AI
Ant Group’s Robbyant released LingBot-VA 2.0, a causal video-action model designed for physical AI systems. It predicts future states at 225 Hz for asynchronous control.
BusinessApple sues OpenAI for allegedly running a "coordinated campaign" to steal trade secrets through poached employees
Apple has filed a lawsuit accusing OpenAI of orchestrating a campaign to poach employees and steal trade secrets related to unreleased products.