China’s AI - News & Prices!
Top News; Fancy Tools; Cheaper/Better Models/Subscriptions at your choice.
AI News
HistoryScroll for more ↓Source-linked briefs
New AI Tools
HistoryScroll for more ↓Source-linked briefs
China’s AI Model Price – You might find one to cut your AI bill!
Official USD quotes · ECB reference FX: 2026-10-15.
Text & code
8 modelsStandard text pricing · USD per 1M tokensAPI prices updated on 300926 · Sources & scope ↗
Peak: weekdays 01–04 & 06–10 UTC, excluding Chinese public holidays.
| MODEL / PROVIDER | USE CASES | CONTEXT | INPUT / 1M | OUTPUT / 1M | Compare |
|---|---|---|---|---|---|
DeepSeek V4.1 FlashDeepSeek · Source ↗ | Text & reasoningCoding · vision | 1M | $0.30 | $1.20 | |
DeepSeek V4 ProDeepSeek · Source ↗ | Text & reasoningCoding | 1M | $1.32 | $3.96 | |
GLM-5.3-FlashZ.AI · Source ↗ | Coding & reasoningMultimodal input | 1M | $0.15 | $0.50 | |
GLM-5.3Z.AI · Source ↗ | Coding & reasoningText input | 1M | $1.40 | $4.40 | |
MiniMax M2.7MiniMax · Source ↗ | CodingText generation | Not checked | $0.30 | $1.20 | |
Coding & reasoningText generation | 1M | $0.30 | $1.20 | ||
Text & reasoningCoding | ≤1M input | $2.00 | $6.00 | ||
Text & codeEfficient tier | ≤1M input | $0.15 | $0.47 |
Image generation
3 quotes| MODEL / PROVIDER | TASK | BILLING SCOPE | PRICE · USD |
|---|---|---|---|
GLM-ImageZ.AI · Source ↗ | Image generation | Published per-image rate; see source for settings | $0.015 / image |
CogView-4Z.AI · Source ↗ | Image generation | Published per-image rate; see source for settings | $0.01 / image |
image-01MiniMax · Source ↗ | Image generation | Published per-image rate | $0.0035 / image |
Rates use the stated units. Features and output settings differ; this is not a quality ranking. The sidebar calculates text tokens only.
Video generation
3 quotes| MODEL / PROVIDER | TASK | BILLING SCOPE | PRICE · USD |
|---|---|---|---|
MiniMax H3MiniMax · Source ↗ | Video generation | 768P · output only | $0.08 / second |
MiniMax H3MiniMax · Source ↗ | Video generation | 2K · output only | $0.13 / second |
CogVideoX-3Z.AI · Source ↗ | Video generation | Per video, not per second; source settings apply | $0.20 / video |
Output price only; input images/video and regeneration can add charges. Per-video and per-second prices are not directly comparable. The sidebar calculates text tokens only.
Audio
3 quotes| MODEL / PROVIDER | TASK | BILLING SCOPE | PRICE · USD |
|---|---|---|---|
Speech RecognitionMiniMax · Source ↗ | Transcription | Audio duration | $0.38 / hour |
speech-2.8-turboMiniMax · Source ↗ | Text to speech | Synchronous synthesis | $0.06 / 1K characters |
speech-2.8-hdMiniMax · Source ↗ | Text to speech | Synchronous synthesis | $0.10 / 1K characters |
Rates use the stated units. Features and output settings differ; this is not a quality ranking. The sidebar calculates text tokens only.
Compare AI Subscription Plans — Which one fits you?
Monthly renewal · introductory offers excluded · ECB reference FX
Most tokens
143.59Mtokens / full quota week · estimated
GLM Coding Max
- 4-week tokens
- 574.36M
- API equivalent
- $1,091.28
- Monthly fee
- $168.00
- Break-even usage
- 15.4%
Break-even at 88.42M total tokens in this scenario.
Potential saving: $923.28For your usage setting · one monthly plan fee
Lowest monthly fee
$18.00/ month, billed monthly
GLM Coding Lite
- 4-week tokens
- 41.03M
- API equivalent
- $77.95
- Break-even usage
- 23.1%
Break-even at 9.47M total tokens in this scenario.
Potential saving: $59.95For your usage setting · one monthly plan fee
Lowest regular monthly fee among the five listed plans.
Best full-use value
6.50×4-week API equivalent / monthly fee
GLM Coding Max
- 4-week tokens
- 574.36M
- API equivalent
- $1,091.28
- Monthly fee
- $168.00
- Break-even usage
- 15.4%
Break-even at 88.42M total tokens in this scenario.
Potential saving: $923.28For your usage setting · one monthly plan fee
How we calculate
Official price snapshot. Token figures are scenario estimates. Four full quota weeks (28 days), compared with one monthly fee. Rankings for tokens and value cover the three GLM plans with published conversion rules.
Input:output 5:1 · 0% cached input · regular peak · 100% usage
Full usage is spread over five usage blocks per week, at least five hours apart, with no opening usage. Both weekly and rolling limits apply. This is not a guaranteed monthly allowance or a quality ranking. MiniMax is excluded from token/value rankings because a numerical quota is not confirmed. How break-even works ↗ · Official credit rules ↗
A lower fee does not mean equivalent model access or quality. MiniMax’s initial discount is excluded; GLM credits depend on model and time of use. Extra usage and taxes may add costs. Sources and held-out candidates ↗
| Plan | Monthly fee | Native allowance / access |
|---|---|---|
| GLM Coding LiteOfficial pricing ↗ | $18.00regular renewal | 10,000 credits / week; 2,000 / 5hGLM-5.3 / GLM-5.3-FlashBoth windows apply; model and time multipliers change consumption. Rules ↗4-week estimate: 41.03M tokens · $77.95 API equivalent · break-even 23.1% |
| GLM Coding ProOfficial pricing ↗ | $80.00regular renewal | 60,000 credits / week; 12,000 / 5hGLM-5.3 / GLM-5.3-FlashNo weekly-to-monthly token guarantee. Promotions excluded. Rules ↗4-week estimate: 246.15M tokens · $467.69 API equivalent · break-even 17.1% |
| GLM Coding MaxOfficial pricing ↗ | $168.00regular renewal | 140,000 credits / week; 28,000 / 5hGLM-5.3 / GLM-5.3-FlashNo weekly-to-monthly token guarantee. Promotions excluded. Rules ↗4-week estimate: 574.36M tokens · $1,091.28 API equivalent · break-even 15.4% |
| MiniMax M Plan GoOfficial pricing ↗ | $22.00regular renewal | Go usage tier; numerical token quota not confirmedM3.1 Flash Preview; image & speechFirst-month offer excluded. Shared usage; no video entitlement. Rules ↗Token equivalent and break-even: unavailable — numerical quota not confirmed. |
| MiniMax M Plan BuildOfficial pricing ↗ | $132.00regular renewal | 7.5× Go usage; not a published monthly token countM3.1 Flash Preview; H3; image & speechFirst-month offer excluded. Shared usage, not one quota per model. Rules ↗Token equivalent and break-even: unavailable — numerical quota not confirmed. |