2026-09-02
Catalog prices: Chinese model IDs on NexoRouter (2 Sep 2026)
Searches like “DeepSeek API price”, “GLM-5.3 cost”, or “Kimi K3 rate” need a list string, not a WeChat rumor. This note lists Chinese IDs that matter for the NexoRouter catalog, with prices captured on 2 September 2026, plus official announcement links for context — without pasting full vendor articles.
How to read this table
- Price source: only nexorouter.com/models on 2 Sep 2026.
- Pending means the ID is listed but has no published rate yet — we do not invent a number.
- Vendor posts explain *what* shipped; they do not replace the gateway price card.
Chat / reasoning (live capture)
| ID | Provider on card | Input / gen. | Cached | Output |
| --- | --- | --- | --- | --- |
| kimi-k3 | Moonshot AI | $3.04 / 1M tokens | $0.30 / 1M tokens | $15.20 / 1M tokens |
| glm-5.3 | Zhipu AI | $1.22 / 1M tokens | $0.30 / 1M tokens | $4.26 / 1M tokens |
| glm-5.3-flash | Zhipu AI | $0.12 / 1M tokens | $0.03 / 1M tokens | $0.43 / 1M tokens |
| glm-5.2 | Zhipu AI | $1.22 / 1M tokens | $0.30 / 1M tokens | $4.26 / 1M tokens |
| glm-5.2-sg | Zhipu AI | $1.59 / 1M tokens | $0.32 / 1M tokens | $5.01 / 1M tokens |
| deepseek-v4-flash-0731 | DeepSeek | $0.24 / 1M tokens | $0.05 / 1M tokens | $0.49 / 1M tokens |
| deepseek-v4-flash-sg | DeepSeek | $0.23 / 1M tokens | $0.05 / 1M tokens | $0.46 / 1M tokens |
| deepseek-v4-pro-sg | DeepSeek | $2.73 / 1M tokens | $0.23 / 1M tokens | $5.47 / 1M tokens |
| Qwen/Qwen3-Max | Alibaba Cloud | $0.35 / 1M tokens | — | $1.40 / 1M tokens |
| qwen3.7-plus | Alibaba Cloud | $1.46 / 1M tokens | $1.82 / 1M tokens | $5.84 / 1M tokens |
| MiniMax-M3 | MiniMax | $2.04 / 1M tokens | $0.41 / 1M tokens | $8.17 / 1M tokens |
| baidu/ernie-4.5-turbo-128k | Baidu | $0.12 / 1M tokens | $0.03 / 1M tokens | $0.49 / 1M tokens |
| doubao-seed-2-0-mini-260215 | ByteDance | $0.03 / 1M tokens | — | $0.28 / 1M tokens |
IDs present with Pending that day (among others): qwen3.8-max, deepseek-v4-pro-0813, deepseek-v4-flash-vision-exp, MiniMax-H3. They appear in the directory; do not cite a price until the string is numeric.
Doubao / MiniMax video & image (official context + card)
ByteDance announced Seedance 2.5 and opened developer APIs via Volcengine / BytePlus in August 2026. On the NexoRouter catalog on 2 Sep 2026:
doubao-seedance-2-0-260128— Input / generation:$0.91 / second(output included)doubao-seedance-2-5-260628— Pending- MiniMax documented H3 open source on 3 Aug 2026; catalog ID
MiniMax-H3was Pending that day.
For Doubao image: doubao-seedream-5-0-pro-260628 showed $0.15 / image.
Tie-back to official announcements (no invented SKUs)
- GLM-5.3 exists in catalog (
glm-5.3,glm-5.3-flash). Z.ai’s 14 Aug 2026 post describes the model; list prices here are NexoRouter’s, not Coding Plan points. - Qwen3.8-Max was announced 3 Aug 2026 and Qwen3.8-Flash-Next on 27 Aug. On NexoRouter,
qwen3.8-maxwas Pending;Qwen/Qwen3-Maxhad a price. Do not mix IDs. - DeepSeek V4 (24 Apr preview, August changelog): use the priced IDs above, not retired names.
Budget without surprises
1. Filter by provider on /models and copy the ID.
2. Multiply estimated tokens by the input/output (and cache) strings.
3. Validate latency and quality in Playground.
4. Read choose models and LatAm latency/currency.
5. If the ID is Pending, do not lock a cost contract until the string is numeric.
Sources
| Date | Source | URL |
| --- | --- | --- |
| 2026-09-02 | Live NexoRouter catalog | https://nexorouter.com/models |
| 2026-08-14 | Z.ai GLM-5.3 | https://z.ai/blog/glm-5.3 |
| 2026-08-03 | Alibaba Qwen3.8-Max | https://www.alibabacloud.com/blog/alibaba-unveils-qwen3-8-max-its-largest-and-most-capable-flagship-model-to-date_603420 |
| 2026-08-27 | Qwen3.8-Flash-Next | https://www.alibabacloud.com/blog/603501 |
| 2026-08-03 | MiniMax H3 OSS | https://www.minimax.io/news/minimax-h3-open-source |
| 2026-04-24 / Aug | DeepSeek V4 + changelog | https://www.deepseek.com/en/news/v4-preview/ · https://api-docs.deepseek.com/updates/ |
Prices change. Before you close a month, reopen /models.