GLM-5.3-Flash

AIHubMix published a dedicated pricing comparison on September 1, 2026 for Z.ai's GLM-5.3-Flash, an efficiency-focused native multimodal model with roughly one million tokens of context that accepts text, image, and video input and targets coding agents, complex reasoning, and long-horizon software engineering. As of S The post positions AIHubMix's offering as a lower current list rate with multi-provider fallback, compared with a direct official Z.ai API or OpenRouter's unified multi-provider routing. It also walks through allocating OpenRouter's 5.5% platform fee proportionally to model usage, giving developers a concrete cost-comp

According to the Substack newsletter, Zhipu AI (operating internationally as Z.ai) shipped GLM-5.3-Flash on August 26, 2026, a 320-billion-parameter mixture-of-experts model that activates roughly 18 billion parameters per token and is described as the first natively multimodal model in the GLM-5 series with a 1-millio The same article reports that Stripe had agreed to acquire OpenRouter the day before Ox Alpha appeared and that Stripe CEO Patrick Collison publicly called the model "very impressive," framing the episode as a milestone for OpenRouter and for non-NVIDIA serving. Z.ai's positioning, as summarized by the newsletter, is t

The Medium explainer from CodeToDeploy confirms that Z.ai released GLM-5.3-Flash on August 26, 2026, as the newest member of the GLM-5 family, designed around delivering frontier-level capabilities without frontier-level inference costs. It specifies 320 billion total parameters with about 18 billion active parameters, The piece frames the launch as a shift in competition from raw model size to intelligence-per-compute, emphasizing agentic coding and dramatically lower inference costs. While it is a third-party recap rather than an official Z.ai post, every specific it reports (release date, parameter counts, context length, license,

A side-by-side comparison by Ampere shows GLM 5.3 Flash (320B total, 18B active) priced at $0.15/$0.50 per million tokens against GLM 5.3 (753B total, 40B active) at $1.40/$4.40, making Flash about one-ninth the cost at normal API pricing. Both models share a 1M-token context window and reasoning plus tool-use capabili The comparison finds GLM 5.3 still leads on raw coding quality, long-horizon engineering tasks, and output speed at roughly 85 tokens per second versus 49 for Flash, while Flash wins on time-to-first-token (1.46s vs 1.57s), visual coding, browser and computer-use agents, and overall cost-performance. GLM 5.3 Flash was

LLM-Stats tracks GLM-5.3-Flash at rank 18 on its composite score, with capability tiers of "Good" in Chat (3 of 117), Tool Calling (7 of 194), and Reasoning (17 of 363), and "Average" in Vision (24 of 208), Coding (32 of 267), Math (38 of 327), Legal (70 of 211), and Finance (86 of 228). Its Quality Tracker showed an i The benchmark scorecard highlights GLM-5.3-Flash's standout GDPval-AA agent economics score of 1773/3000 (rank 2, evaluated by Artificial Analysis) and a CharXiv-R reasoning score of 0.89/1 (rank 8) reported by Z.ai. Terminal-Bench 2.1 and other agentic coding evaluations are listed with rankings and methodology detail

Z.ai released GLM-5.3-Flash on August 26, 2026, as a 320-billion-parameter mixture-of-experts model that activates only 18 billion parameters per token and runs natively in FP8. According to the technical deep dive, it is the first natively multimodal model in the GLM-5 series, supporting text, image, and video input w The launch revealed that the previously anonymous "Ox Alpha" model, which had been quietly topping OpenRouter usage charts for about a week before release, was in fact GLM-5.3-Flash running a preview build. The deep dive notes the vLLM serving recipe is already public and that the model beats GLM-5.2 across six coding

Fello AI's coverage confirms Z.ai's August 26, 2026 release of GLM 5.3 Flash as a 320B-parameter MoE activating 18B parameters per token, built on a newly trained base rather than a post-train of GLM 5.2. It is the first natively multimodal model in the GLM-5 series, pairs a 1M-token context window with a hybrid sparse Pricing is reported at $0.15 per million input tokens and $0.50 output, compared with $1.40 and $4.40 for GLM 5.3, and the article notes that the flagship GLM 5.3's 744B open-weight release promised "in about two weeks" had not materialized, with no GLM 5.3 repository yet on Z.ai's Hugging Face organization. The articl