Baichuan API Pricing (November 2026)
Baichuan API prices run from $0.146 to $14.88 per million tokens across 9 current models. The cheapest is Baichuan4 Air at $0.146 input and $0.146 output per million tokens. The flagship, Baichuan4, costs $14.88 input and $14.88 output per million tokens. Every rate below was read on Baichuan's own pricing page, on the date shown in its row.
| Model | Input /M | Cache read /M | Output /M | Batch in /M | Batch out /M | Context | Status | Checked |
|---|---|---|---|---|---|---|---|---|
| Baichuan4 | $14.88 | — | $14.88 | — | — | 32K | Flagship | 2026-10-29 |
| Baichuan4 Turbo | $2.23 | — | $2.23 | — | — | 32K | Current | 2026-10-29 |
| Baichuan4 Air | $0.146 | — | $0.146 | — | — | 32K | Current | 2026-10-29 |
| Baichuan3-Turbo (128K) | $3.57 | — | $3.57 | — | — | 128K | Current | 2026-10-29 |
| Baichuan-M3 | $1.49 | — | $4.46 | — | — | 32K | Current | 2026-10-29 |
| Baichuan3-Turbo | $1.79 | — | $1.79 | — | — | 32K | Current | 2026-10-29 |
| Baichuan-M3-Plus | $0.744 | — | $1.34 | — | — | 32K | Current | 2026-10-29 |
| Baichuan-M2-Plus | $1.49 | — | $4.46 | — | — | 32K | Current | 2026-10-29 |
| Baichuan-M2 | $0.298 | — | $2.98 | — | — | 32K | Current | 2026-10-29 |
| Baichuan Text Embedding | $0.0744 | — | Free | — | — | documented elsewhere | Current | 2026-10-29 |
| Baichuan2-53B | $1.49 | — | $1.49 | — | — | 32K | Legacy | 2026-10-29 |
| Baichuan2-Turbo | $1.19 | — | $1.19 | — | — | 32K | Legacy | 2026-10-29 |
US dollars per million tokens, list rates on Baichuan's own API. Cache read is the price of a cached input token. Batch is the asynchronous API, where the vendor offers one. The date in each row is when we last read it on the vendor's pricing page; it links to that page.
Baichuan API monthly cost estimates
What light use (10M input + 2M output tokens a month), medium use (100M input + 20M output tokens a month), heavy use (1B input + 200M output tokens a month) costs on each current model, at list price with no caching and no batch discount.
| Model | Input / output per 1M | Light · 10M + 2M | Medium · 100M + 20M | Heavy · 1B + 200M |
|---|---|---|---|---|
| Baichuan4 Air | $0.146 / $0.146 | $1.75 | $17.50 | $175 |
| Baichuan-M2 | $0.298 / $2.98 | $8.93 | $89.28 | $893 |
| Baichuan-M3-Plus | $0.744 / $1.34 | $10.12 | $101 | $1,012 |
| Baichuan-M3, Baichuan-M2-Plus | $1.49 / $4.46 | $23.81 | $238 | $2,381 |
| Baichuan3-Turbo | $1.79 / $1.79 | $21.43 | $214 | $2,143 |
| Baichuan4 Turbo | $2.23 / $2.23 | $26.78 | $268 | $2,678 |
| Baichuan3-Turbo (128K) | $3.57 / $3.57 | $42.85 | $429 | $4,285 |
| Baichuan4 | $14.88 / $14.88 | $179 | $1,785 | $17,855 |
Prompt caching and the Batch API cut these figures; the next sections show by how much. For your own token counts, use the LLM API cost calculator.
Baichuan API price changes and releases
Releases and corporate moves, with the 2025–2026 pivot to medical AI front and center. Sourced from Baichuan press, 36Kr, and Wikipedia — verified at publication.
platform.baichuan-ai.com — the managed API for the full Baichuan catalog. Most general SKUs use a unified per-token rate (same price for input and output); prices are published in CNY and converted here at ¥6.7209/USD.
OPEN-WEIGHT · SELF-HOSTBaichuan open-sources select models — the Baichuan-M3 medical model and the earlier Baichuan2 (7B / 13B) weights are published for self-hosting.
CONSUMER · APPBaixiaoying (百小应) — Baichuan's consumer assistant app for end users in China. Separate from the developer API.
Alibaba (Qwen)
Broad open-weight catalog with 119-language coverage and many sizes. Baichuan can't match the breadth, but leads on specialised medical models.
COST DISRUPTORDeepSeek
MIT open weights at the lowest cost-per-token for general reasoning. Baichuan's general SKUs are pricier; its differentiation is healthcare, not cost.
SIX-TIGERS PEERZhipu (GLM)
GLM family with free tiers and strong agentic models. A fellow "Six Little Tigers" startup; Zhipu is broader, Baichuan is medical-deep.
SIX-TIGERS PEERMoonshot (Kimi)
1T-param open-weight MoE and long-context heritage. Another Beijing "Tiger"; Moonshot chases frontier scale where Baichuan chases clinical depth.
PRIMARY RIVALOpenAI
Frontier benchmarks + ecosystem (GPT-5). Baichuan claims its medical M3 outperforms GPT-5.2-class models on clinical tasks specifically — a narrow but pointed challenge.
GLM ARCHITECTZhipu (Z.ai / GLM)
The Beijing lab behind the GLM models, also operating internationally as Z.ai. Spun out of Tsinghua University in 2019, Zhipu ships strong coding/agentic models, multiple free tiers, and open weights — and was the first of China's "Six Tigers" to pursue an IPO. It also sits on the US Entity List.
Baichuan API pricing FAQ
Practical questions about Baichuan's models, medical focus, and pricing.
Q · 01 Which Baichuan model should I start with?
+
$2.11/M unified) is the value pick; Baichuan4 Air ($0.14/M) is the budget option. For clinical/medical tasks, use the M-series — Baichuan-M3 ($1.49/$4.46). The premium Baichuan4 flagship is $14.09/M. See the use case picker. Q · 02 What is the medical M-series?
+
Q · 03 Why is Baichuan4 so expensive, and what's 'unified pricing'?
+
$14.09/M both ways, a premium tier; Baichuan4 Turbo and Air are far cheaper for the same general use. Q · 04 Is there a surcharge on the medical models?
+
Q · 05 Where is my data stored, and are models open-weight?
+
Q · 06 How does Baichuan compare to the other Chinese labs?
+
Baichuan Intelligence (百川智能) was founded on April 10, 2023 in Beijing by Wang Xiaochuan, the former CEO of search engine Sogou. It is one of China's so-called "Six Little Tigers" of AI, alongside peers like Zhipu, Moonshot, and MiniMax, and has raised over $1 billion from backers including Alibaba, Tencent, and Xiaomi at a valuation around $2.7 billion.
Baichuan's defining move is its pivot to healthcare AI. After launching general-purpose models (the Baichuan3 and Baichuan4 lines in 2024), the company refocused from early 2025 on medical models — the M-series. In January 2026 it released Baichuan-M3, an open-source medical model reported to top several authoritative medical evaluations and to outperform frontier general models, and even average human physicians, on some clinical tasks.
The general lineup remains: Baichuan4 (a premium flagship at a unified $14.09/M), the cheaper Baichuan4 Turbo, the budget MoE Baichuan4 Air, and the long-context Baichuan3-Turbo 128K. A notable quirk is Baichuan's unified pricing on most general SKUs — the same per-token rate for input and output — plus a separate medical-search surcharge on some M-series calls.
Baichuan open-sources select models (Baichuan-M3 and the earlier Baichuan2 7B/13B), while keeping its top general SKUs API-only. Founder Wang Xiaochuan has said the company targets a 2027 IPO in China.
The trade-offs: models are text-only and Chinese-language-focused, the company is China-based (data stored under Chinese rules on the direct API), and US export-control questions apply. Versus broad peers like Qwen or DeepSeek, Baichuan's edge is narrow but deep: specialised, benchmark-leading medical AI rather than a general frontier race.