June 17, 2026

Z AI’s GLM-5.2 leads Intelligence Index with score of 51

Z ai has released its GLM-5.2 model, which has attained the top position as the leading open weights model on the Artificial Analysis Intelligence Index, scoring 51 points. This model, while the same size as its predecessor GLM-5.1, shows significant improvements, particularly in scientific reasoning and long-horizon tasks, with enhanced performance on various evaluations. The GLM-5.2 is also competitively priced and widely available through both Z ai’s API and multiple third-party providers, ensuring broad accessibility for developers. Its advancements place it alongside proprietary models, making it a notable contender in the artificial intelligence landscape.

Z ai: Z.ai is a Chinese AI company formerly known as Zhipu AI that develops large language models under the GLM series. It recently released GLM-5.2, an open-weights model focused on long-horizon agentic and coding tasks with expanded context capabilities. The company makes the model available through its first-party API and multiple third-party inference providers.
Baseten: Baseten is a third-party infrastructure company that hosts GLM-5.2 among other models for production use. It streamlines deployment and scaling for developers adopting new open-weights releases. The service supports the wider AI tooling ecosystem.
GLM-5.2: GLM-5.2 is Z.ai’s latest open-weights large language model in the GLM-5 line, designed for advanced reasoning, scientific tasks, and agentic performance. It features a significantly larger context window than its predecessor and is positioned as a leading open model on key intelligence benchmarks. The model is released under an MIT license and accessible via Z.ai’s API as well as various third-party platforms.
GPT-5.5: GPT-5.5 is a proprietary frontier model from OpenAI known for high-level reasoning and performance on complex, long-horizon tasks. It serves as a benchmark reference for comparing open-weights models on real-world agentic metrics. The model represents ongoing proprietary advancements in intelligence and coding capabilities.
nebiusai: Nebius AI is an infrastructure provider that integrates and serves cutting-edge models including GLM-5.2 via its platform. It supports the deployment of open-weights models for broader adoption. The company contributes to the ecosystem of accessible AI compute resources.
DeepInfra: DeepInfra is a third-party AI inference platform that hosts and serves various large language models including GLM-5.2. It enables developers to access new models through scalable API endpoints shortly after release. The platform supports multiple providers in the open-weights ecosystem.
GMI_cloud: GMI Cloud serves as a third-party inference provider for GLM-5.2 and similar models. It enables scalable access to high-performance AI systems for users beyond direct channels. The platform focuses on reliable model deployment.
Kimi K2.6: Kimi K2.6 is a leading open-weights model from the Kimi series, frequently benchmarked against peers in intelligence, reasoning, and efficiency metrics. It provides competitive performance in scientific and agentic tasks while maintaining strong context handling. The model is part of ongoing advancements in accessible high-capability AI systems.
MiniMax-M3: MiniMax-M3 is an open-weights AI model from MiniMax emphasizing coding, agentic capabilities, and multimodal features with a large context window. It serves as a direct comparator in recent benchmark evaluations for intelligence and real-world task performance. The model is positioned as a frontier-level open offering in the current AI landscape.
ParasailAI: Parasail AI operates as a third-party host for large language models, making GLM-5.2 available through its services. It facilitates easy access and experimentation with newly released models. The platform is part of the growing network of inference providers.
novita_labs: Novita Labs is a third-party provider offering inference services for advanced AI models such as GLM-5.2. It helps expand accessibility of new models beyond first-party channels. The service focuses on reliable hosting for developers and researchers.
siliconflow: SiliconFlow is a third-party AI platform that provides hosting and API access to models like GLM-5.2. It supports developers working with open-weights releases across various benchmarks and use cases. The service enhances distribution of advanced AI tools.
FireworksAI_HQ: Fireworks AI is a third-party platform providing fast inference for models including GLM-5.2. It helps make newly announced models immediately usable through optimized endpoints. The company plays a key role in accelerating access to frontier open models.
DeepSeek V4 Pro: DeepSeek V4 Pro is a high-performance variant in DeepSeek’s V4 series of large language models, noted for strong results in agentic coding and reasoning benchmarks. It competes directly with other leading open-weights models on metrics like intelligence index and GDPval evaluations. The model is available through API and emphasizes cost-effective long-context processing.

Model Leadership: GLM-5.2 establishes itself as the top open-weights model on the Artificial Analysis Intelligence Index while remaining competitive with proprietary systems on agentic benchmarks.
Availability Expansion: The model launches simultaneously on Z.ai’s API and a wide array of third-party inference providers, broadening immediate access for developers.
Capability Improvements: GLM-5.2 demonstrates notable gains in scientific reasoning and long-horizon agentic tasks compared to prior versions, with an expanded context window enabling more complex workflows.

AI tools built by Emerald Force

Built and supported by Emerald Force.

You might also like