Last verified
ENTERPRISE BY APPLICATIONLEGAL WORKFLOWSSAME RATE AS FLASHX1M CONTEXTIMAGE + VIDEO IN

GLM-5.3-FlashX-Legal API Pricing

GLM-5.3-FlashX-Legal is GLM-5.3-FlashX trained further for legal work: research, contract review and drafting legal documents. Z.ai prices it exactly like FlashX - $0.37/M input, $0.075/M cached input and $1.25/M output - but API access goes only to selected enterprise customers by application. It shares FlashX's spec: 1M context, image, video and file input, thinking always on. Z.ai published no release date; it first appeared in October 2026.

Input - per 1M tokens
$0.37/M
Access by application same as FlashX
Output - per 1M tokens
$1.25/M
Thinking always on thinking billed here
Cached input - per 1M tokens
$0.075/M
Cache storage free for now -80%
Effective - agentic blend
$0.218/M
92/8 split - 82% cache hit rate
§ 01 / TERMINAL

GLM-5.3-FlashX-Legal API cost calculator.

Calculator pre-loaded with GLM-5.3-FlashX-Legal's rates, which only an approved enterprise account can use. Legal work is input-heavy - contracts, case files, statutes - so the input rate and the cache hit rate move the bill most.

$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
Open full calculator (all models · share URL · CSV) →
§ 02 / SCENARIOS

GLM-5.3-FlashX-Legal cost on real workloads.

CONTRACT REVIEW

Review a 40-page contract

$0.00759/contract
30,000 in - 3,000 out~13,170 units/$100
DUE DILIGENCE

Data room pass

$0.077/pass
500,000 in - 10,000 out~1,306 units/$100
RESEARCH

Case law memo

$0.016/memo
80,000 in - 5,000 out~6,061 units/$100
DRAFTING

First draft of a filing

$0.010/draft
20,000 in - 6,000 out~9,938 units/$100
§ 03 / TOKENIZER

GLM-5.3-FlashX-Legal token counter and cost.

What's a token →
Calibrated · measured on the vendor's tokenizer · 2026-06-10 Auto-counts as you type

Counts use a chars-per-token calibration measured on the vendor's own published tokenizer (zai-org/GLM-5, 2026-06-10). English prose is typically within a few percent; code and non-Latin scripts tokenize heavier. For billing-exact counts use the vendor's count-tokens API.

Characters 366
Words 56
Tokens (estimated) 70 tokens
Cost as input · uncached $0.00003 USD
Cost as output · uncached $0.00009 USD
Cost as cached input $0.00001 USD
§ 04 / SHELF

GLM-5.3-FlashX-Legal vs similar models.

All models →
Model Input /M Output /M Effective blended Context Best for
GLM-5.3-FlashX-Legal Current $0.37 cache $0.075 $1.25 $0.218 agentic 92/8 1M Legal research, contract review and drafting
GLM-5.3-FlashX $0.37 cache $0.075 $1.25 $0.218 same 1M The same base model, open to everyone
GLM-5.3-Flash $0.15 cache $0.03 $0.50 $0.0875 −2.5× 1M The slower Flash tier at a lower rate
GLM-5.3 $1.40 cache $0.26 $4.40 $0.78 +3.6× 1M The full-size flagship, text only
MiniMax M3.1 Flash $0.15 cache $0.015 $0.60 $0.0842 −2.6× 1M Cheap 1M-context agents with image input
DeepSeek V4.1 Flash $0.30 cache $0.006 $1.20 $0.15 −31% 1M Cheap 1M-context agent loops with images

GLM-5.3-FlashX-Legal API pricing FAQ.

What GLM-5.3-FlashX-Legal costs, how to get access, and how it differs from the GLM-5.3-FlashX it is built on.

Q · 01 How much does GLM-5.3-FlashX-Legal cost? +
$0.37/M input, $0.075/M cached input and $1.25/M output on Z.ai's International card, the same as GLM-5.3-FlashX. That is $0.22/M on our 92/8 agentic blend with 82% cache hits. Storing cached input is free for a limited time, a separate charge this site does not model.
Q · 02 Who can use it? +
Selected enterprise customers. Z.ai's model page says API access is granted by application, through its legal consultation form; there is no self-serve key for it. Z.ai also offers it inside AutoClaw Legal Assistant, a workspace with plugins for Word and WPS. Because access is gated, this site keeps it off the public leaderboards.
Q · 03 How is it different from GLM-5.3-FlashX? +
It is the same 320B-parameter model with legal-domain training on top: mid-training on legal data, then reinforcement learning in tens of thousands of task environments modelled on legal work. Z.ai reports a 19.2% all-criteria pass rate on Harvey's LAB benchmark for legal agents, 12.5 points above the base model; that is Z.ai's figure, not ours. Price, context and inputs are the same.
Q · 04 When was it released? +
Z.ai has not said. Its release notes do not mention the model. The Internet Archive has the model page without it on October 12, 2026 and with it on October 21, and it was on the pricing card by October 23. This page dates it to the day we first verified the price.
Q · 05 What does it need to work well? +
Z.ai recommends running it in an agent connected to legal MCP services, which supply laws, cases and company data and edit Word files with tracked changes. Those services sit outside the token price on this page.
Y
Reviewed by Yaroslav Vikhariev Founder · AI//COST · Prices checked by hand against docs.z.ai · Last verified Oct 23, 2026
Methodology Report a correction More by Y.V.