GLM-4.7-Flash

GLM-4.7-Flash is designed as a pragmatic, high-performance alternative to massive proprietary systems, specifically optimized for coding tasks and agentic workflows. Unlike its larger sibling, which utilizes a complex Mixture-of-Experts architecture, this model employs a streamlined, dense 30-billion parameter structure. This design choice ensures predictable inference behavior and consistent latency, making it a reliable workhorse for teams that require robust code generation without the hardware complexities often associated with larger, sparse models.

Built for versatility and ease of deployment, the model supports fine-tuning and is compatible with local execution environments, allowing developers to integrate it into their own infrastructure. Its dense architecture simplifies quantization, enabling efficient operation on consumer-grade hardware while maintaining strong performance across programming scenarios and reasoning benchmarks. By balancing capability with resource efficiency, it offers a practical path for teams looking to scale agentic workflows and repository-level analysis without the overhead of larger, more demanding systems.

NovitaAIzai-org/glm-4.7-flashglm-flash

Open provider page Provider docs

Quick Info

Provider: NovitaAI
Model key: zai-org/glm-4.7-flash
Release date: Jan 19, 2026
Last updated: Jan 19, 2026
Knowledge cutoff: 2025-04
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $0.07
Output token cost: $0.40

Limits

Output tokens: 128,000 tokens
Context window: 200,000 tokens

Transparent token rates

Compare GLM-4.7-Flash pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GLM-4.7-Flash

NovitaAI

Official sourceComparison

Compare GLM-4.7 and GLM-4.7-Flash by tier: flagship quality vs scalable cost, with real benchmarks, speed data, and Novita pricing., Post a Comment

NovitaAI

CoverageBenchmark

GLM-4.7 Flash packs 31B parameters and an MIT license with free API access, helping you test ideas and ship tools on a tiny budget.

NovitaAI

Coverage

The 30B model achieves 59.2% on SWE-bench Verified while running at 80+ tokens per second on MacBooks.

NovitaAI

CoverageRelease Notes

GLM-4.7-Flash is a new member of the GLM 4.7 family and targets developers who want strong coding and reasoning performance in a model that is practical to...

Hugging Face

CoverageBenchmark

Zhipu AI (Z.ai) officially open-sourced GLM-4.7-Flash, a "Hybrid Thinking" model positioned as the strongest performer in the 30B class. It uses a 30B-A3B Mixture-of-Experts architecture, activating roughly 3B parameters per task to balance resource usage with processing power. Across key benchmarks, GLM-4.7-Flash reac The release emphasizes developer-friendly local deployment, with vLLM and SGLang already supporting the model on main (configurable via tensor-parallel-size, speculative-config, and the EAGLE algorithm) and Hugging Face transformers enabling direct invocation. The model targets agent applications in local or private-cl