GLM-5.3-Flash
GLM-5.3-Flash is the first natively multimodal model in Z.ai's GLM-5 series, with 320B total and 18B activated parameters and a hybrid of linear and sparse attention. Its weights are released under the MIT licence.
Released
August 26, 2026
Type
multimodal
License
open-weight
MIT
Context
—
Z.ai states a “1M-token” context window.
Pricing
Input
$0.15 / 1M tokens
Output
$0.50 / 1M tokens
List price on the developer's own pages, checked . Cached input $0.03; cached input storage limited-time free. Faster GLM-5.3-FlashX (about 200 tokens/s): $0.37 input / $0.075 cached / $1.25 output.
Capabilities
Architecture
Parameters: 320B total, 18B activated
Sources
Every figure on this page was checked against these pages on . A figure we could not source is left out.
- Z.ai docs: New Released (release notes) (opens in a new tab)
- Z.ai docs: GLM-5.3-Flash/FlashX (opens in a new tab)
- Z.ai docs: Pricing (opens in a new tab)
- Hugging Face: zai-org/GLM-5.3-Flash model card (opens in a new tab)
Spotted something out of date? Tell us.
Links
In the news: GLM-5.3-Flash and Z.ai (Zhipu AI)
MarkTechPost
Z.ai GLM-Legal Beats Claude Fable 5 and Kimi K3 on Harvey LAB
Z.ai has released GLM-Legal, a legal AI model designed for tasks such as contract review, evidence analysis, and compliance research. The model is built on Z.ai's GLM-5.3-Flash architecture and has achieved a 19.2% all-pass rate on Harvey’s Legal Agent Benchmark (LAB). This rate is nearly three times higher than its base model's 6.7% pass rate. GLM-Legal is a domain-specialized version of Z.ai's GLM-5.3-Flash model, which is natively multimodal and can process text, images, video, and files, outputting text. The base model features 320 billion total parameters with 18 billion active parameters, utilizing a mix of sparse and linear attention to reduce serving costs for long contexts. The model underwent a two-stage training process. The first stage involved mid-training on expanded legal corpora to enhance reasoning from facts to conclusions. The second stage included post-training with reinforcement learning across tens of thousands of legal task environments, with the agent harness co-evolving with the model. Performance metrics indicate GLM-Legal achieved a 19.17% Harvey LAB task pass rate, ranking seventh out of 79 models on Vals AI, at a cost of $0.15 per test. While its performance on Vals AI’s Legal Research Bench was 41.83%, below its base model's 45.19%, Z.ai states GLM-Legal offers frontier-class legal agent scores at approximately 1/128 the cost of Claude Fable 5. Access to GLM-Legal is currently limited, with its API available only through application. Public weights are not provided, and the model runs exclusively on Z.ai’s hosted API.
MarkTechPost
Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks
Cantina Security, in collaboration with Yeta Labs, has introduced apex-flash-1, an open-weights artificial intelligence model specifically engineered for vulnerability research. This model is a reinforcement learning fine-tune of Z.ai's GLM-5.3-Flash and has been made available on Hugging Face under the permissive MIT license. The apex-flash-1 model is deployable using popular frameworks such as vLLM, SGLang, or Transformers, though its BF16 version requires approximately 640 GB of GPU memory. The model boasts a total of 312.3 billion parameters, as indicated in its Hugging Face safetensors metadata. Its base, GLM-5.3-Flash, is a Mixture-of-Experts model featuring 18 billion active parameters. Cantina employed a training methodology that included GRPO, a rank-256 LoRA, and selective full-parameter training. The training dataset comprised 150 tasks derived from 50 real-world vulnerability cases, with each case presented in three distinct variants: guided whitebox, focused whitebox, and focused blackbox. The vulnerability types covered include authorization, identity, and scope flaws, which constitute 72% of the cases. Additionally, accounting and numerical precision bugs account for 18% of the dataset, with the remaining cases involving time validation, business rules, and Server-Side Request Forgery (SSRF) vulnerabilities. According to the model card published on Hugging Face, the reinforcement learning rollouts were conducted within the Codex agent harness, utilizing environments that closely mimic production software and protocols. In an evaluation using 60 tasks drawn from 20 distinct, held-out vulnerability cases, apex-flash-1 demonstrated significant capabilities. The model successfully solved 40 out of these 60 tasks, achieving a pass@1 rate of 66.7%, at an estimated cost of $2.38 per run. For comparative analysis, the base GLM-5.3-Flash model solved 36 out of 60 tasks (60.0% pass@1) at a cost of approximately $4.56. When benchmarked against Anthropic's Claude 3 Opus 5 High, apex-flash-1's performance was competitive, with Opus solving 43 out of 60 tasks (71.7% pass@1) but at a substantially higher cost of about $74.68 per run. This indicates that apex-flash-1 achieved a solved task cost of roughly $0.06, compared to $1.74 for Claude Opus. These performance figures and costs are based on internal benchmarks reported by Cantina Security. Cantina positions apex-flash-1 not as a standalone orchestrator, but as a specialized worker model designed to be integrated into larger AI systems. Its intended applications include code reading, tool utilization, exploit development, and verification tasks. An experimental variant, apex-flash-1-abliterated, which features modified refusal behavior, has also been developed but was not separately evaluated. The company's rationale is that security defenders require highly capable AI models to effectively identify and address vulnerabilities.
MarkTechPost
GLM-5.3-Flash vs Qwen3.8-Flash-Next: Two Chinese AI Labs Independently Converge on the Same Model Architecture
Two leading Chinese artificial intelligence labs, Z.ai and Alibaba's Qwen team, independently developed and released frontier open-weight models with strikingly similar architectures within a single week. Z.ai launched GLM-5.3-Flash, a 320 billion parameter multimodal Mixture-of-Experts (MoE) model featuring 18 billion active parameters. Concurrently, Alibaba's Qwen team unveiled Qwen3.8-Flash-Next, a 125 billion parameter model with 6 billion active parameters, which serves as a preview of the upcoming Qwen4 architecture. Despite their independent development, the configurations of these two models are nearly identical, indicating a potential convergence in cutting-edge AI model design. Both models employ a 3:1 hybrid attention mechanism, combining linear and full attention. They also utilize a compressed indexer to manage context, capped at 2048 tokens, and widen the residual stream into four gated branches. Furthermore, both models were trained using the Muon optimizer, with fused parameter matrices split prior to orthogonalization. GLM-5.3-Flash is the first natively multimodal model in the GLM-5 series, released under the MIT license on Hugging Face. Z.ai tested it anonymously as Ox Alpha on OpenRouter, where it quickly became the most popular model of the week. This model was trained on a substantial 30 trillion token multimodal corpus and supports a context window of 1 million tokens. Z.ai claims that GLM-5.3-Flash outperforms its predecessor, GLM-5.2, across various benchmarks at one-tenth the cost, while achieving performance comparable to Claude Opus 4.8 on coding and agentic tasks. Its listed pricing is $0.15 per million input tokens and $0.50 per million output tokens. Qwen3.8-Flash-Next fulfills a similar role to Qwen3-Next in the Qwen3.5 release, acting as an early public preview of the next generation of Qwen architecture. The model card specifies a main model of 125 billion parameters, supplemented by an additional 51 billion n-gram embedding table, with 6 billion parameters activated per token. Its native context length is 262,144 tokens, which can be extended to 1 million tokens using YaRN. The Qwen team reported that the training of Qwen3.8-Flash-Next required approximately one-ninth of the compute resources needed for Qwen3.7-Plus. The technical report accompanying this release is titled “On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability.” The shared architectural choices suggest a consensus among leading AI research labs on effective strategies for building powerful and efficient large language models, particularly in the realm of multimodal capabilities and context handling. The only notable point of disagreement among the described configurations is a single architectural detail, and one lab has reportedly dissented from the prevailing design choices.
MarkTechPost
Z.ai Releases GLM-5.3-Flash: A 320B-A18B Natively Multimodal MoE With a 1M-Token Context
Z.ai has released GLM-5.3-Flash, a new natively multimodal model that represents a significant advancement in the GLM-5 series. This model is characterized by its Mixture-of-Experts (MoE) architecture, featuring a total of 320 billion parameters with 18 billion active per token. A key feature is its exceptionally large context window, extending to 1,048,576 tokens, which allows for the processing of extensive amounts of information in a single input. GLM-5.3-Flash also boasts native multimodal capabilities, enabling it to process both image and video inputs. The model has been released under an MIT license, with its weights made available on Hugging Face, facilitating broader access and adoption by the research and development community. According to Z.ai's reports, GLM-5.3-Flash demonstrates superior performance compared to its predecessor, GLM-5.2, across various benchmarks and real-world applications. Notably, it achieves this enhanced capability at approximately one-tenth the price of previous models. In internal coding benchmarks, GLM-5.3-Flash performs competitively, achieving a score within half a point of Claude Opus 4.8. The model initially ran anonymously for its first week on platforms like OpenCode and OpenRouter, utilizing domestically produced Chinese AI chips for its operations. This release positions GLM-5.3-Flash as a cost-effective and powerful tool for a range of AI tasks. The deployment of GLM-5.3-Flash is structured across two primary tracks, ensuring accessibility for different user needs. The model's weights are publicly available on Hugging Face under the permissive MIT license, allowing organizations to self-host and integrate it into their infrastructure. Additionally, a priced and operational hosted API is already serving user requests, providing a more accessible entry point for those who may not have the resources for self-hosting. The default FP8 checkpoint requires approximately 306 GiB of storage before considering the KV cache, and the current vLLM implementation supports NVIDIA Hopper architecture and newer GPUs. This hardware requirement means that self-hosting is realistically within reach for mid-size to large organizations possessing at least an 8-GPU node or equivalent processing power, as well as AI-native startups that rent GPU capacity. For entities unable to self-host due to hardware or resource constraints, the API offers an economical solution where the focus shifts from hardware management to operational costs. Industries identified as having an immediate fit for GLM-5.3-Flash include software development and devtools, IT and Business Process Outsourcing (BPO) automation, financial services and insurance for document operations, enterprise business intelligence, back-office knowledge work, e-commerce, and any team responsible for shipping user interfaces at scale. Potential applications are diverse, ranging from developing repo-scale coding agents and terminal/browser-use agents to analyzing million-token logs and contracts, performing UI regression checks from screenshots, and enabling complex reasoning over spreadsheets, presentations, and dashboards that would typically require an Optical Character Recognition (OCR) to text pipeline.