GLM-5.3 API Launches at $1.40 Input, $4.40 Output Per Million Tokens
Z.ai launches GLM-5.3 via API at $1.40 input and $4.40 output per million tokens, offering a lower-cost option for frontier coding and agent workloads.
Summary
On Aug. 19, 2026, Chinese startup Z.ai made GLM-5.3 available through its API, following a debut in which its cyber capabilities reportedly uncovered a previously undetected Cursor vulnerability. GLM Coding Plan subscribers remain limited to an OpenAI Chat Completions-compatible protocol. Z.ai plans to release the model weights, but has not specified a date or license.
Pricing remains level with GLM-5.2 at $1.40 per million input tokens and $4.40 per million output tokens. Cached input costs $0.26, with storage temporarily free. VentureBeat’s comparison of one million input plus one million output tokens totals $5.80, against $8 for Grok 4.6, $18 for Kimi K3, $30 for Claude Opus 5 and $35 for GPT-5.6 Sol. GLM-5.3 is not cheapest: GPT-5.6 Luna costs $0.20 input and $1.20 output, while Gemini 3.7 Flash costs $0.75 and $3.75 through Dec. 31, 2026.
Z.ai claims stronger coding and long-horizon agent performance without a higher posted token rate. Artificial Analysis scored GLM-5.3 at 60, seven points above GLM-5.2 and tied with Kimi K3 as the leading open weights model. However, its estimated Intelligence Index task cost rose from about $0.44 to $0.68 because GLM-5.3 is more verbose, showing that workload bills still depend on output volume, input mix and caching.
Positives
- $1.40 input and $4.40 output pricing matches GLM-5.2 despite Z.ai’s claims of stronger coding and long-horizon agent performance.
- Artificial Analysis scored GLM-5.3 at 60, seven points above GLM-5.2 and level with Kimi K3 among open weights models.
- $5.80 for one million input plus one million output tokens undercuts Grok 4.6, Kimi K3, Claude Opus 5 and GPT-5.6 Sol.
- $0.26 cached input and temporarily free cached-input storage can reduce API costs for repeated context.
- GLM-5.3’s API release lets developers integrate the model into agents and applications immediately.
Risks & concerns
- $0.68 per Intelligence Index task exceeds GLM-5.2’s roughly $0.44 because GLM-5.3 generates more tokens.
- Z.ai has provided no release date or licensing terms for GLM-5.3’s planned open weights.
- GLM Coding Plan subscribers are currently restricted to the OpenAI Chat Completions-compatible protocol.
- GPT-5.6 Luna and Gemini 3.7 Flash remain cheaper than GLM-5.3 at their listed rates.
- $5.80 is only a simplified comparison, since actual bills vary with token consumption, input-output mix and caching.