GLM-5.3
GLM-5.3 is Z.AI's flagship reasoning model, built for complex software engineering, long-horizon agent tasks, and security work such as vulnerability discovery.
input · usd / 1m tok
$1.40
output · usd / 1m tok
$4.40
context window
1mtokens
modalities · in → out
takes text in, gives text out
pricing
base api ratesusd per 1m tokens
- input
- $1.40
- output
- $4.40
- cache read
- $0.26
providers
Different companies host GLM-5.3. Impossibl routes your request intelligently and handles failover automatically when a provider is down.
- Z.AI
- Fireworks
specifications
takes text in, gives text outzai/glm-5.3
- creatorZ.AI
- model typeText generation
- context window1,000,000 tokens
- maximum outputnot published
- input
takes text in, gives text out
- outputtext
- tool callingnot verified
- releasedaug 14, 2026
- endpoint
/v1/chat/completions
unknown means the public catalog does not establish that fact. a model's documented capabilities can differ from what a particular api supports. see the model documentation for integration details.
try it · your prompt, this model, live
# responses API
import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["IMPOSSIBL_API_KEY"],
base_url="https://api.impossibl.com/v1",
)
resp = client.responses.create(
model="zai/glm-5.3",
input="",
)
print(resp.output_text)
edit the prompt, then run
benchmarks
intelligence index
44.9#7 of 90 models on the gateway
- coding index74.8#13 of 60
- agentic index53.4#4 of 49
- omniscience14.3#20 of 86
- gpqa91.7%#25 of 88
- humanity’s last exam42.3%#22 of 89
- terminal-bench v2.183.9%#15 of 57
- GPQA91.7%
- Terminal-Bench v2.183.9%
- AA-LCR79.7%
- AutomationBench Partial62.2%
- SciCode59.0%
- Tau Banking50.3%
- Multilingual LCR48.3%
- Humanity’s Last Exam42.3%
- Terminal-Bench v4.041.9%
- CritPt19.1%
- GDP.pdf All-pass11.2%
- Intelligence Index44.86
- Omniscience Index14.30
- GDPval Elo1667.27
- coding index (API)74.8
- agentic index (API)53.4
- intelligence index (API)44.9
each configuration is listed separately; these evaluations describe the tested model configuration, and performance and costs can differ on impossibl. unmeasured scores are omitted and api values keep their source units. last checked 2026-09-11.
compare GLM-5.3 with other models
frontier labs and open weights, best-ranked firstusd / 1m tok · in / out
- anthropic/claude-fable-5-1$10.00 / $50.00compare →
- openai/gpt-6-astra$10.00 / $50.00compare →
- meta/muse-spark-1.3$1.25 / $4.25compare →
- moonshotai/kimi-k3$3.00 / $15.00compare →
frequently asked questions
What is GLM-5.3?
GLM-5.3 is Z.AI's flagship reasoning model, built for complex software engineering, long-horizon agent tasks, and security work such as vulnerability discovery. It succeeds GLM-5.2 with a large gain on Z.AI's own coding bench and supports long text inputs with tool calling.
How much does GLM-5.3 cost?
GLM-5.3 costs $1.40 per million input tokens and $4.40 per million output tokens through the Impossibl API. Cache reads cost $0.26 per million tokens. All rates are in USD.
What is the context length of GLM-5.3?
GLM-5.3 has a 1,000,000-token context window through the Impossibl API.
What inputs and outputs does GLM-5.3 support?
GLM-5.3 accepts text as input. GLM-5.3 returns text.
How do I use GLM-5.3 through an API?
GLM-5.3 is available through the Impossibl API. Use zai/glm-5.3 as the model identifier with /v1/chat/completions. The quickstart explains how to create an API key and send your first request.
When was GLM-5.3 released?
Z.AI released GLM-5.3 on August 14, 2026.
this page describes GLM-5.3 as listed by the impossibl public api. prices and availability come from that catalog, refreshed here every five minutes. this is an impossibl offering, not a comparison of independent hosting-provider quotes. model descriptions and release dates, where available, use the creator sources linked under the answers above.
view the source catalog ↗model as markdown ↗report a correction →