DEV.co
Open-Source LLM · zai-org

GLM-5-FP8

GLM-5-FP8 is a 753B-parameter open-source LLM from zai-org, released April 2026. It uses sparse mixture-of-experts architecture with FP8 quantization to reduce deployment cost. The model is designed for complex reasoning, coding, agentic tasks, and tool use. MIT-licensed, ungated, with 2M+ downloads. Supports multiple serving frameworks (vLLM, SGLang, KTransformers, Transformers, xLLM). Context length is not specified in the card.

Source: HuggingFace — huggingface.co/zai-org/GLM-5-FP8
753.9B
Parameters
mit
License (OSI-approved)
Unknown
Context (tokens)
2M
Downloads (30d)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Developerzai-org
Parameters753.9B
Context windowUnknown
Licensemit — OSI-approved
Modality / tasktext-generation
Gated on HuggingFaceNo
Downloads2M
Likes181
Last updated2026-04-05
Sourcezai-org/GLM-5-FP8

What GLM-5-FP8 is

GLM-5-FP8 scales to 744B total parameters (40B active) trained on 28.5T tokens. Incorporates DeepSeek Sparse Attention (DSA) to reduce memory footprint. FP8 quantization applied. Post-training uses 'slime', an asynchronous RL infrastructure. Supports tool calling (glm47 parser) and reasoning modes (glm45 parser). Multi-modal capacity not mentioned in card. Benchmarks show competitive performance on AIME, IMO, SWE-bench, agentic tasks (Terminal-Bench, CyberGym, BrowseComp), and reasoning (HLE). Multilingual: English and Chinese.

Quickstart

Run GLM-5-FP8 locally

Load the open weights with 🤗 Transformers and generate — the same model, self-hosted.

quickstart.pypython
from transformers import pipelinepipe = pipeline("text-generation", model="zai-org/GLM-5-FP8")out = pipe("Explain retrieval-augmented generation in one sentence.",           max_new_tokens=128)print(out[0]["generated_text"])

Swap in vLLM or Ollama for production-grade serving. DEV.co can stand up the inference stack.

Deployment

How you'd run it

A typical self-hosted path — open weights, an inference server, your application.

DEV.co builds each layer — from GPU infrastructure to the application.

Best use cases

Software Engineering Automation

SWE-bench Verified 77.8% and SWE-bench Multilingual 73.3% scores indicate strong code generation and debugging capability. Suitable for automated codebase analysis, PR review, or test generation at scale.

Agentic Systems & Tool Orchestration

Terminal-Bench 2.0 (56.2%), CyberGym (43.2%), and BrowseComp (75.9% with context management) show strong multi-step task execution. Applicable for autonomous agents, workflow automation, and system-level problem solving.

Complex Reasoning & Long-Context Analysis

HLE 30.5 and IMOAnswerBench 82.5 demonstrate strong reasoning. Supports 131k+ token generation. Suitable for document synthesis, multi-turn analysis, and complex problem decomposition in specialized domains.

Running & fine-tuning it

ESTIMATE: FP8 quantization with 744B parameters ≈ 372GB raw model weight (744B × 0.5 bytes/param FP8). Effective working memory (KV cache, activations) for typical serving: ~600–800 GB total GPU VRAM for batch size 1–4. vLLM recipe specifies 8× GPUs (e.g., 8× H100 80GB ≈ 640GB) at 85% utilization. Activation memory and context length not disclosed; 131k token context window noted in evals likely requires > 1TB aggregate memory. Exact specifications require testing.

Card does not discuss LoRA, QLoRA, or instruction-tuning capability. Post-training used proprietary 'slime' RL infrastructure; reproducibility and fine-tuning API unknown. Standard Transformers library support (v0.5.4+) suggests PyTorch/HuggingFace compatibility, but fine-tuning guides are not referenced. Requires review of GitHub repo or model card for supervised fine-tuning and adapter method feasibility.

When to avoid it — and what to weigh

  • Real-time, Ultra-low-latency Requirements — 744B parameters (40B active) still requires significant GPU resources. Inference latency likely measured in hundreds of ms even with speculative decoding. Unsuitable for sub-100ms SLAs or high-frequency inference.
  • Resource-Constrained or Edge Deployments — FP8 + MoE quantization reduces footprint but still demands 8+ high-end GPUs per the vLLM recipe. Not feasible on CPU, mobile, or constrained edge devices without further quantization (INT4/INT2).
  • Multimodal Input (Vision, Audio) is Required — Card describes text-generation pipeline only. No mention of image, video, or audio modalities. Use closed-source alternatives (GPT-5, Claude, Gemini) if multimodal input is mandatory.
  • Guaranteed Factual Accuracy or Hallucination-Free Output — Standard LLM behavior; no guardrails or retrieval-augmented generation mentioned. Requires integration with external knowledge bases or fact-checking pipelines for factual-critical applications.

License & commercial use

MIT License. Permissive OSI-approved license permitting commercial use, modification, and redistribution with attribution. No restrictions on deployment model (closed-source, SaaS, on-premise, etc.).

MIT is a permissive OSI license that explicitly allows commercial use. No gating, no academic-only clause, no clause requiring derived works to be open-source. Commercial applications, proprietary integrations, and SaaS deployment are permitted. No support guarantees, SLAs, or commercial licensing provided by zai-org. Commercial users should review the LICENSE file on the GitHub repo and conduct their own compliance review if bundling or modifying.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationStrong
License clarityClear
Deployment complexityHigh
DEV.co fitStrong
Assessment confidenceHigh
Security considerations

Standard LLM security considerations apply: prompt injection, jailbreaking, token leakage, and model extraction are possible. No safety fine-tuning, content filtering, or guardrails mentioned in the card. Inference on untrusted inputs requires containerization and rate-limiting. FP8 quantization does not inherently improve or degrade security. Access to sparse attention implementation (DSA) is open-source, enabling security researchers to audit but also potential adversaries to optimize attacks. Sensitive data should not be stored in long context windows without encryption.

Alternatives to consider

DeepSeek-V3.2

Similar dense-to-sparse scale. Open-source. Comparable on HLE (25.1 vs 30.5), SWE-bench (73.1 vs 77.8), and agentic tasks (BrowseComp 67.6 vs 75.9). May offer different trade-offs in latency or context handling; review benchmarks for your use case.

Claude Opus 4.5 (Proprietary)

Closed-source, commercial API. Stronger on some benchmarks (HLE-with-Tools 43.4*, SWE-bench Verified 80.9, CyberGym 50.6). If maximum reasoning and tool-use performance is critical and cost/latency are secondary, consider Claude. No self-hosting or IP control.

Llama 3.3 70B or 405B (Open-source)

Smaller parameter count, likely lower inference cost and latency. Llama license (LLAMA2 variant) is permissive for most commercial uses. Benchmarks on reasoning and agentic tasks typically lag GLM-5; use if VRAM/cost is the limiting factor and acceptable performance floor is lower.

Software development agency

Ship GLM-5-FP8 with senior software developers

GLM-5-FP8 is production-ready for software engineering automation, long-context reasoning, and multi-step agentic workflows. MIT-licensed and self-hostable. Review hardware requirements, benchmark performance for your domain, and test deployment on your target GPU cluster. Contact our AI engineering team to evaluate fit and design a custom fine-tuning or evaluation pipeline.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

GLM-5-FP8 FAQ

Can I use GLM-5-FP8 commercially?
Yes. MIT license permits commercial use without restriction. You may deploy it as a closed-source service, integrate it into proprietary products, or modify and redistribute it under MIT terms. No commercial licensing agreement or permission from zai-org is required, though you must retain the MIT license notice.
What GPU hardware do I need to serve GLM-5-FP8?
Minimum 8× high-end GPUs (H100 80GB, A100, RTX 6000) to reach reasonable latency per the vLLM recipe. Aggregate VRAM ~600–800 GB. Smaller deployments (single GPU or < 200GB) require further quantization (INT4/INT2) and may incur unacceptable latency degradation. Test on target hardware before production.
What is the context length of GLM-5-FP8?
Not explicitly stated in the card. Evals use contexts up to 202k tokens (HLE-with-tools) and 250-minute timeouts (CyberGym). The model likely supports at least 128k–256k tokens, but verify in the GitHub repo or test yourself before assuming a specific limit.
Does GLM-5-FP8 support vision or multimodal input?
No. The card describes text-generation only. No image, video, audio, or multimodal capability is mentioned. If you need vision, use alternatives like GPT-4V, Claude 3.5, or Gemini 3 Pro, or combine GLM-5 with a separate vision model.

Software developers & web developers for hire

DEV.co is a software development agency delivering custom software development services to companies building on open source. Our software developers and web developers design, integrate, and ship production systems — spanning web development, APIs, AI, data, and cloud. If GLM-5-FP8 is part of your open-source llms roadmap, our team can implement, customize, migrate, and maintain it.

Deploy GLM-5-FP8 for Complex Reasoning and Agentic AI

GLM-5-FP8 is production-ready for software engineering automation, long-context reasoning, and multi-step agentic workflows. MIT-licensed and self-hostable. Review hardware requirements, benchmark performance for your domain, and test deployment on your target GPU cluster. Contact our AI engineering team to evaluate fit and design a custom fine-tuning or evaluation pipeline.