Z.ai Releases GLM-5.2 with 1M Token Context Window at $1.40/$4.40 per Million
Z.ai has released GLM-5.2, a model designed for long-horizon engineering tasks with a 1 million token context window. The model is priced at $1.40 per million input tokens and $4.40 per million output tokens, and was released on June 16, 2025.
GLM-5.2 at a glance
Z.ai Releases GLM-5.2 with 1M Token Context Window
Z.ai has released GLM-5.2, a model with a 1 million token context window designed for project-level engineering tasks. The model is priced at $1.40 per million input tokens and $4.40 per million output tokens.
Technical Specifications
- Context window: 1 million tokens
- Input pricing: $1.40 per 1M tokens
- Output pricing: $4.40 per 1M tokens
- Release date: June 16, 2025
- Modalities: Text in/out
Claimed Capabilities
According to Z.ai, GLM-5.2 is positioned as their "flagship model for the era of long-horizon tasks." The company claims the model can:
- Handle project-level engineering context
- Execute long-running tasks with improved reliability
- Follow engineering standards consistently
- Complete full development workflows from requirements to multi-platform deployment
The model is currently hosted exclusively through OpenRouter, with all requests forwarded directly to Z.ai's infrastructure.
Pricing Context
At $1.40/$4.40 per million tokens, GLM-5.2's pricing positions it in the mid-range for large context window models. For comparison:
- Anthropic's Claude 3.5 Sonnet (200K context): $3/$15 per 1M tokens
- OpenAI's GPT-4o (128K context): $2.50/$10 per 1M tokens
- Google's Gemini 1.5 Pro (2M context): $1.25/$5 per 1M tokens
OpenRouter notes that effective pricing can be 60-80% lower than list prices when prompt caching is applied for repeated context.
What This Means
GLM-5.2 enters a competitive market for long-context models, where context window size alone no longer differentiates offerings. The 1M token window matches several existing models, while models like Gemini 1.5 Pro already offer 2M tokens at comparable pricing. The real test will be whether Z.ai's claimed advantages in long-horizon task execution and engineering workflow completion translate to measurable performance improvements in production use cases. Without published benchmark scores or independent verification of the model's capabilities, its market position remains uncertain.
Related Articles
Mistral releases final Le Chonk, a trillion-parameter open-weight model it calls best for coding and cybersecurity
Mistral has released the finalized version of Le Chonk, a trillion-parameter open-weight model that is now freely available. The company claims it is the world's best open-weight model for automated coding and cybersecurity, matching or surpassing models from DeepSeek and Z.ai.
NVIDIA and MIT team releases Open-H3-Regenerate-2K, regenerating MiniMax H3 video drafts at up to native 4K
The Efficient-Large-Model team, made up of NVIDIA and MIT researchers, released Open-H3-Regenerate-2K weights on Hugging Face. Built on MiniMax H3, it regenerates low-resolution video drafts at 2560×1440 and up to native 3840×2160. The team reports a 3.9x speedup on a single B200 and a higher UniPercept score than the official H3-Regenerate-2K API.
Pika releases H3 Flex, an open-weight video model that generates synchronized audio in six sampling steps
Pika Labs has published Pika H3 Flex, a post-trained H3 model that generates video with synchronized audio in six sampling steps and accepts up to 12 reference inputs. Open weights run locally at native 480p and 768p, while Pika's hosted API adds 1080p, 2K and 4K output.
Edge0 releases Audio8-TTS-Realistic, a 0.6B open-weight voice cloning TTS model under Apache 2.0
Edge0 released Audio8-TTS-Realistic, a 0.6B open-weight zero-shot voice cloning text-to-speech model under Apache 2.0. It supports 23 languages and adds a final GRPO stage driven by a learned realism critic.
Comments
Loading...