Chinese AI startup Z.ai (formerly Zhipu AI) has officially announced the immediate release of GLM-5.2, a 753-billion parameter open-weights large language model (LLM) engineered specifically to dominate “long-horizon” autonomous coding and engineering tasks.
Available immediately on Hugging Face, the Z.ai API, and more than 20 third-party coding environments, the model boasts a highly stable 1-million-token context window. Comprehensive enterprise subscription tiers start at a highly competitive $12.60 per month.
The power of the MIT license
In excellent news for cost- and security-conscious businesses, Z.ai has released GLM-5.2’s core weights under an unrestricted MIT open-source license. This allows enterprises to download the model freely, customize or fine-tune it to their liking, and run it locally or via virtual machines for only the cost of their compute and electricity.
This option is increasingly appealing for global enterprises as state-of-the-art proprietary models face an uncertain and potentially interrupted regulatory future. Following last week’s Trump Administration export control directive prohibiting foreign nationals from using Anthropic’s new Claude Fable 5 model, Anthropic responded by taking the models in question entirely offline for all users. GLM-5.2 provides a highly capable path to host frontier-level AI locally, entirely bypassing geographic fencing and commercial limitations.
IndexShare and selectable thinking modes
Under the hood, GLM-5.2 operates with 753 billion parameters and introduces a major architectural optimization called “IndexShare.” In standard massive language models, recalculating attention mechanisms across long documents is computationally exorbitant. IndexShare solves this by reusing the identical indexer across every four sparse attention layers. At the maximum 1-million-token context length, this single innovation reduces per-token compute FLOPs by a massive 2.9 times.
The model also features an upgraded Multi-Token Prediction (MTP) layer for speculative decoding, which boosts accepted token length by up to 20% during inference.
Additionally, Z.ai has implemented flexible, selectable “Thinking Modes,” allowing users to toggle the model’s reasoning effort based on their needs:
- “Max” mode: Pushes the limits of logical problem-solving to peak intelligence, utilizing nearly 85k output tokens per task.
- “High” mode: Strikes a careful balance between high-end performance and latency-sensitive token efficiency, effectively halving the required token output.
Benchmark dominance
On industry-standard third-party benchmark tests, GLM-5.2 performs above most open-source flagship models, including DeepSeek v4, and scores near or above its closed-weights rivals, OpenAI’s GPT-5.5 and Anthropic’s Claude Opus 4.8. The model particularly shines in agentic tool use and long-horizon software engineering tasks.
In software engineering evaluations, GLM-5.2 scored 62.1 on SWE-bench Pro, decisively beating GPT-5.5 (58.6) and its own predecessor, GLM-5.1 (58.4). On FrontierSWE, designed to test long-horizon task completion, GLM-5.2 hit 74.4%, surpassing GPT-5.5 (72.6%) and finishing in a near-tie with Claude Opus 4.8 (75.1%).
On tool-usage evaluations like MCP-Atlas, GLM-5.2 achieved a 77.0, outscoring GPT-5.5 (75.3) and performing just shy of Claude Opus 4.8 (77.8). When equipped with external tools on Humanity’s Last Exam, GLM-5.2 reached a score of 54.7, coming out ahead of GPT-5.5 (52.2) and tracking closely behind Claude Opus 4.8 (57.9).
In extended, multi-hour engineering workloads, GLM-5.2 consistently topped GPT-5.5, scoring 34.3% against GPT-5.5’s 25.0% on PostTrainBench, and 13.0% against GPT-5.5’s 12.0% on SWE-Marathon.
While GLM-5.2 trails Claude Opus 4.8 and GPT-5.5 slightly on raw Terminal-Bench 2.1 scores (81.0 versus 85.0 and 84.0, respectively), it significantly outscores Google’s Gemini 3.1 Pro (74.0). Beyond traditional coding metrics, GLM-5.2 took an impressive first place on the crowdsourced design task benchmark Design Arena, beating out even the state-of-the-art Claude Fable 5 with an ELO score of 1360.
Aggressive pricing
To operationalize the model, Z.ai launched the GLM Coding Plan, aiming squarely at developer workflows rather than simple chat interfaces. The plan offers out-of-the-box support for third-party agentic coding harnesses and tools including Claude Code, OpenClaw, Cline, Kilo Code, Crush, and Factory.
When billed annually, the Coding Plan pricing tiers are highly disruptive:
- Lite: $12.60 per month ($151.20 per year starting in the 2nd year) – geared toward lightweight iteration on small repositories.
- Pro: $50.40 per month – for day-to-day development on mid-sized repositories, offering 5x the usage allowance of the Lite plan.
- Max: $112.00 per month – for heavy workloads, offering 20x the Lite usage and dedicated resources during peak hours.
For developers integrating the raw model, Z.ai’s API pricing undercuts its Western rivals significantly while matching the exact rates of the previous GLM-5.1 generation. GLM-5.2 API access is priced at $1.40 per million input tokens and $4.40 per million output tokens (a total of $5.80).
The stark contrast between open-weights innovators and proprietary Western labs has sparked intense discussion. On X (formerly Twitter), prolific AI observer Lisan al Gaib (@scaling01) argued that frontier labs are “absolutely scamming you on API pricing,” noting that while massive open models operate profitably without relying on the newest Blackwell chips, proprietary labs demand heavy premiums and are likely sitting at 90%+ margins. For comparison:
- OpenAI GPT-5.5: Input $5.00 | Output $30.00 (Total: $35.00)
- Anthropic Claude Opus 4.8: Input $5.00 | Output $25.00 (Total: $30.00)
- Anthropic Claude Fable 5 / Mythos 5: Input $10.00 | Output $50.00 (Total: $60.00)
To further optimize costs for long-context workloads, Z.ai offers a cached input rate of just $0.26 per million tokens, alongside a limited-time offer for free cached input storage.
The developer reaction to the release has been immediate and overwhelmingly positive. The team behind Kilo Code confirmed day-one integration, posting on X: “GLM-5.2 runs in Kilo Code on day one. The 1M context window and Max effort mode are both live. Point your config at it and go!”
Open-source coding environment Cline IDE echoed this sentiment, highlighting the economic and performance advantages:
“GLM-5.2 is the first open-weights model to cross 80% on Terminal-Bench, and beats every other open model available. It also beats Gemini, making it a frontier-level model for a fraction of the cost. Open weights is back. This model is a game changer. Available in Cline now!”
Similarly, rival open-source desktop agent Eigent AI tested the model’s capabilities on complex agentic workflows—such as researching 30 companies across 6 sectors of the AI infrastructure stack, structuring it into JSON, and building an interactive HTML report—noting that GLM-5.2 pulls ahead significantly in task planning and long-horizon execution.
Source: venturebeat.com
