Zhipu AI's GLM-4.5 is a versatile text-to-text large language model designed for diverse natural language processing tasks. Featuring a 128,000-token context window, it supports understanding and generating very long-form text with high coherence and contextual awareness.
GLM-4.5 aims to close the gap by integrating diverse capabilities within a single framework. Across 12 benchmarks spanning agentic tasks (3), reasoning (7), and coding (2), GLM-4.5 achieves an overall third-place ranking. Its variant, GLM-4.5 Air, ranks sixth, showing competitive yet slightly lower results compared to top models from OpenAI, Anthropic, Google DeepMind, xAI, Alibaba, Moonshot, and DeepSeek.
.png)
Vs. Claude 4 Sonnet: GLM-4.5 shows close performance in agentic coding and reasoning tasks but still has optimization potential compared to Claude Sonnet 4, which achieves excellent coding success and state-of-the-art reasoning outputs, often favored for autonomous multi-feature app development.
Vs. OpenAI GPT-4.5: GLM-4.5 ranks overall competitive with top models including GPT-4.5 in reasoning and agent benchmarks, though GPT-4.5 generally leads in raw task accuracy on some professional benchmarks like MMLU and AIME.
Vs. Qwen3-Coder and Kimi K2: GLM-4.5 dominates Qwen3-Coder with an 80.8% success rate and wins 53.9% of tasks against Kimi K2, highlighting its superior coding and agentic capabilities for complex programming scenarios.
Vs. Gemini 2.5 Pro: GLM-4.5 holds its ground on reasoning and coding benchmarks but detailed exact percentages are less public; Gemini 2.5 stands stronger in certain coding and reasoning tests but GLM-4.5 balances large context and agentic tools well.
The full GLM-4.5 model requires substantial computational resources and GPU memory, potentially limiting deployment feasibility for organizations with constrained infrastructure. The lighter GLM-4.5 Air variant addresses this but with fewer active parameters and slightly reduced capabilities.
Measured on the AI/ML API gateway across live traffic. Figures are medians unless stated. Last updated 2026-10-12.
import requests r = requests.post( "https://api.aimlapi.com/v1/chat/completions", headers={"Authorization": "Bearer " + AIMLAPI_KEY}, json={ "model": "zhipu/glm-4.5", "messages": [ { "role": "user", "content": "Hello!" } ] }, ) print(r.json())
const r = await fetch("https://api.aimlapi.com/v1/chat/completions", { method: "POST", headers: { Authorization: `Bearer ${process.env.AIMLAPI_KEY}`, "Content-Type": "application/json", }, body: JSON.stringify({ "model": "zhipu/glm-4.5", "messages": [ { "role": "user", "content": "Hello!" } ] }), }); console.log(await r.json());
curl -X POST https://api.aimlapi.com/v1/chat/completions \ -H "Authorization: Bearer $AIMLAPI_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"zhipu/glm-4.5","messages":[{"role":"user","content":"Hello!"}]}'
OpenAI-compatible — swap the base URL and it works with your existing SDK.
| Type | Price |
|---|---|
| Input | |
| Output | |
| Cached input |
| Benchmark | Score | What it measures | Source | Retrieved |
|---|---|---|---|---|
| Terminal-Bench | 37.5% | Autonomous shell/terminal task completion | Source | July 8, 2026 |
| SWE-bench Verified | 64.2% | Resolving verified real GitHub issues | Source | July 8, 2026 |
| Intelligence | 12.8 | Composite score across standardised reasoning, knowledge and problem-solving evaluations, measured independently by Artificial Analysis | Source | September 12, 2026 |
| Math | 73.7 | Composite score across standardised mathematics evaluations, measured independently by Artificial Analysis | Source | September 12, 2026 |
| Model | Input | Output | Context | Best for |
|---|---|---|---|---|
GLM-4.5 This page | Reasoning + agents | |||
| Reasoning + agents | ||||
| Balanced coding + agents | ||||
| Long-context, multimodal & agentic workflows | ||||
| Reasoning + agents |
GLM-4.5 has a 128,000 tokens context window and can return up to 98,000 tokens.
GLM-4.5 takes text as input and returns text.
Use zhipu/glm-4.5 as the model id. Requests go to https://api.aimlapi.com/v1/chat/completions.
GLM-4.5 became available on September 9, 2025.
GLM-4.5 is priced at input $0.78 / 1M tokens, output $2.86 / 1M tokens, cached input $0.143 / 1M tokens.
Yes, GLM-4.5 can stream responses as they are generated.
GLM-4.5 was built by Zhipu AI.
