FLAGSHIP GLM1M CONTEXT128K OUTPUTDEEP THINKINGTOOL STREAMINGTEXT + CODE
GLM-5.2 API Pricing
GLM-5.2 is Z.AI's current flagship text model for long-horizon coding and project-scale engineering. The live vendor table lists $1.4/M input and $4.4/M output, with cached input at $0.26/M. The model doc lists 1M context and 128K max output.
Input - per 1M tokens
$1.40/M
Source Z.AI flat
Output - per 1M tokens
$4.40/M
Context 1M flat
Cached input - per 1M tokens
$0.26/M
Storage limited-time free -81%
Effective - agentic blend
$0.78/M
92/8 split - 82% cache
§ 01 / TERMINAL
GLM-5.2 API cost calculator.
Live calculator pre-loaded with current GLM-5.2 rates. Tweak spend, output mix, or cache assumptions for 1M-context coding and agentic workloads.
$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
§ 02 / SCENARIOS
GLM-5.2 cost on real workloads.
CODING AGENT
Repo implementation
$0.023/task
CODE REVIEW
Pull request review
$0.014/review
RAG
Knowledge base answer
$0.00859/query
CHATBOT
Support assistant
$0.0038/turn
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 614
Words 88
Tokens (estimated) 117 tokens
Cost as input · uncached $0.00016 USD
Cost as output · uncached $0.00051 USD
Cost as cached input $0.00003 USD
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| GLM-5.2 Current | $1.40 cache $0.26 | $4.40 | $0.78 agentic 92/8 | 1M | 1M-context engineering |
| GLM-5.1 | $1.40 cache $0.26 | $4.40 | $0.78 same | 200K | Predecessor flagship |
| GLM-5 | $1.00 cache $0.20 | $3.20 | $0.572 −27% | 200K | Cheaper GLM-5 base |
| GLM-5 Turbo | $1.20 cache $0.24 | $4.00 | $0.70 deprecated · −10% | 200K | Delisted August 2026 - kept for reference |
| GLM-4.7 | $0.60 cache $0.11 | $2.20 | $0.358 −2.2× | 200K | Budget coding agent |
| DeepSeek V4 Pro | $1.32 cache $0.044 | $3.96 | $0.569 −27% | 1M | Low-cost reasoning |
| Qwen3.8 Max | $2.00 | $6.00 | $2.32 +3.0× | 1M | Alibaba flagship |
GLM-5.2 has published scores on 15 benchmarks we track; its best placing is #50 of 99 on ARC-AGI-1. 2 of these scores come from an independent leaderboard; the rest are the lab's own reports.
| Benchmark | Score | Rank | Leader | Source |
|---|---|---|---|---|
| ARC-AGI-1 Reasoning | 77% ARC Prize verified, best result | #50 of 99 | 6 tied · 98.5% | Official leaderboard ARC Prize Foundation |
| ARC-AGI-2 Reasoning | 22.8% ARC Prize verified, best result | #55 of 105 | GPT-6 Astra · 95% | Official leaderboard ARC Prize Foundation |
| Terminal-Bench 2 Agents and computer use · Terminal-Bench 2.1 | 81% as reported | #11 of 17 | DeepSeek V4.1 Flash · 90.6% | Vendor-reported Z.ai (Zhipu) |
| Humanity's Last Exam Expert knowledge · With tools | 54.7% with tools | #15 of 20 | Claude Fable 5.2 · 74.2% | Vendor-reported Z.ai (Zhipu) |
| FrontierSWE Coding · Version not stated | 67.5% as reported | #4 of 5 | Kimi K3 · 81.2% | Vendor-reported Z.ai (Zhipu) |
| AutomationBench Agents and computer use · Main set (as reported) | 26.2% v1.0.6 | #19 of 22 | MiniMax M3.1 Flash · 62.8% | Vendor-reported Z.ai (Zhipu) |
| CyberGym Cybersecurity | 77.2% as reported | #7 of 8 | MiMo-V2.6-Flash · 95.1% | Vendor-reported Z.ai (Zhipu) |
| NL2Repo-Bench Coding | 48.9% as reported | #7 of 8 | DeepSeek V4.1 Flash · 68.2% | Vendor-reported Z.ai (Zhipu) |
| DeepSWE Coding | 46.2% as reported | #24 of 27 | Gemini 4 Argon · 77.9% | Vendor-reported Z.ai (Zhipu) |
| Toolathlon Agents and computer use | 59.9% as reported | #8 of 9 | MiMo-V2.6-Pro · 76.9% | Vendor-reported Z.ai (Zhipu) |
| Agents' Last Exam Agents and computer use · CLI split | 23.8% CLI split | #6 of 6 | Kimi K3.1 Pro · 44.2% | Vendor-reported Z.ai (Zhipu) |
| ExploitBench Cybersecurity | 24.4% as reported | #7 of 7 | GPT-6 Astra · 100% | Vendor-reported Z.ai (Zhipu) |
| PostTrainBench AI research | 31.7% as reported | #6 of 6 | Gemini 4 Argon · 45.3% | Vendor-reported Z.ai (Zhipu) |
| ProgramBench Coding | 9.5% almost-solved rate | #7 of 7 | Kimi K3.1 Pro · 31.5% | Vendor-reported Z.ai (Zhipu) |
| Terminal-Bench 3.0 Agents and computer use | 4.6% as reported | #9 of 9 | DeepSeek V4.1 Flash · 30% | Vendor-reported Z.ai (Zhipu) |
GLM-5.2 API pricing FAQ.
Straight answers for teams estimating GLM-5.2 API bills, migration from GLM-5.1, and the effect of caching on long-context engineering workloads.
Q · 01 What is GLM-5.2 priced at? +
Z.AI lists
GLM-5.2 at $1.4/M input, $0.26/M cached input, and $4.4/M output. Cached input storage is marked as limited-time free on the pricing page.Q · 02 How is the effective price calculated? +
The headline effective tile uses the site standard agentic blend:
92% input, 8% output, and 82% input cache hits. For GLM-5.2, that lands at about $0.78/M blended tokens.Q · 03 Is GLM-5.2 more expensive than GLM-5.1? +
No. The current Z.AI pricing table lists GLM-5.2 and GLM-5.1 at the same API rates:
$1.4/M input, $0.26/M cached input, and $4.4/M output.Q · 04 What changed versus GLM-5.1? +
The model page positions GLM-5.2 as the long-horizon flagship with
1M context, 128K maximum output, stronger coding benchmarks, deep thinking, function calling, structured output, context caching, MCP, and tool streaming support.Q · 05 Is prompt caching priced separately? +
Yes. The pricing page lists cached input at
$0.26/M. Cached input storage is currently shown as limited-time free, so this page models cache-hit input but not storage charges.Q · 06 How accurate is the tokenizer estimate? +
The widget uses a GLM-family planning estimate of
3.85 characters per token. Exact billing can vary by language, tool calls, hidden reasoning tokens, and cache boundaries.