Published API pricing per 1M tokens
Source: each model's own OpenRouter listing, fetched late August 2026. Sorted cheapest first on blended cost at a 1:3 input/output ratio.
| Model | Input $/1M | Output $/1M | Blended $/1M | Context (tokens) | Providers on OpenRouter |
|---|---|---|---|---|---|
| DeepSeek V4 Pro 0813 | $0.25 | $0.75 | $0.375 | 1,000,000 | 3 |
| Kimi K3 | $0.60 | $3.00 | $1.05 | 262,144 | 4 |
| GLM-5.3 | $1.40 | $4.40 | $2.65 | 1,048,576 | 4 |
| Qwen3.8-Max | $2.00 | $6.00 | $3.00 | 1,000,000 | 1 |
| Claude Opus 4.8 | $15.00 | $75.00 | $45.00 | 200,000 | 2 |
How blended cost is calculated
Blended cost assumes a 1:3 input/output token ratio (the workload pattern most API providers publish against). At 1:1 GLM-5.3 blends to $2.90 and Claude Opus 4.8 to $45.00 — the gap stays large in either direction.
Ready to test the workflow?
Create account & add creditsAnchors built on "cheapest" or "budget" claims
Any anchor built on "cheapest" or "budget" pointed at GLM-5.3 is a claim the table above disproves. DeepSeek V4 Pro 0813 is roughly 5.6× cheaper on input and 5.9× on output. The defensive position for GLM-5.3 is open-weight-adjacent reasoning depth with native multimodal input, not the cheapest line item.
Frequently asked questions
How much does GLM-5.3 cost per token?
$1.40 per 1M input tokens and $4.40 per 1M output tokens on the Z.ai / OpenRouter listing as of late August 2026. Blended at 1:3 input/output, $2.65 per 1M tokens.
Is GLM-5.3 the cheapest flagship-tier model?
No. DeepSeek V4 Pro 0813 is roughly 5.6× cheaper on input and 5.9× on output. GLM-5.3's position is open-weight-adjacent reasoning depth with native multimodal input — not the cheapest line item.
Where is the live GLM-5.3 model page?
The canonical model page with current OneInfer pricing, capabilities, and availability is /models/zai-org/GLM-5.3. This page is a focused facet of that entity, not a replacement for it.
How should I treat benchmark or price claims?
Check each claim’s provenance label and observed date. Vendor-reported and independently verified numbers are shown as separate evidence classes on this hub.
Put GLM-5.3 to work
Fund a controlled evaluation, start with a prepared prompt, and measure quality and cost on your own workload.