Qwen3.8-Max API TWD Cost 2026: AA Index 45.4, Flagship

Query date: 2026-09-20 (Taipei time). Benchmark data comes from the live Artificial Analysis website page. Pricing follows the USD unit prices shown on the BazaarLink models page, and the exchange rate is 1 USD = NT$ 31.5.

Qwen3.8-Max is Alibaba's latest flagship model: a 2.4-trillion-total-parameter MoE architecture (about 95 billion parameters active per inference), a 1-million-token context, and deep reasoning. After its listing on BazaarLink, you can call it directly with the model name qwen3.8-max.

Bottom line up front: Qwen3.8-Max scores 45.4 on the Artificial Analysis Intelligence Index. The order among the three is Qwen3.8-Max 45.4, GLM-5.3 44.8, and Kimi K3 43.6, and its price is lower than both of these flagships. Its generation speed is the slowest of the three (39.0 tokens/second), so keep this in mind when evaluating latency-sensitive scenarios.

Want to call Qwen3.8-Max and pay in TWD with a unified invoice that carries your tax ID? BazaarLink offers an OpenAI-compatible route with the same model unit prices, plus a 10% top-up service fee, addressing the expense-reporting pain point for Taiwan companies. See the complete guide to AI API expense reporting for Taiwan companies for details.

Third-party benchmark: Artificial Analysis Intelligence Index

(Data from the Artificial Analysis website's qwen3-8-max model page, query date 2026-09-20)

Qwen3.8-Max is currently one of the higher-scoring models in the Chinese camp. The order among the three is Qwen3.8-Max 45.4, GLM-5.3 44.8, Kimi K3 43.6. This score has been revised before: Artificial Analysis's first measurement was 53, but it later found that some tests ran on an endpoint with intermittent issues, and after re-running, the score recovered to 56. When citing third-party benchmark figures, keep this revision history in mind: a score is not permanently fixed once a single test is done.

Price comparison (USD ↔ NT$)

Cache-hit input $0.25/MTok (NT$ 7.9). The prices in this table match the BazaarLink models page, with no model markup. For the latest unit prices, see BazaarLink Models.

Scenario estimate: deep-reasoning batch tasks (NT$/month)

Assumptions: 300 reasoning tasks per day (math, logic, code analysis), each with 3,000 input tokens and 1,200 output tokens.

• Monthly tasks: 300 × 30 = 9,000 calls
• Monthly input tokens: 9,000 × 3,000 = 27,000,000 (27 MTok)
• Monthly output tokens: 9,000 × 1,200 = 10,800,000 (10.8 MTok)

Cost:

• Input: 27 × NT$ 63 = NT$ 1,701
• Output: 10.8 × NT$ 189 = NT$ 2,041
• Total: NT$ 3,742/month (≈ USD 119)

→ The same workload on Kimi K3: NT$ 7,830/month (1.1× more expensive, and 1.8 points lower on the intelligence index); on Opus 5: NT$ 12,780/month (2.4× more expensive). For batch tasks that are highly cost-sensitive and can tolerate slower speed, Qwen3.8-Max is the cheapest of the three Chinese-camp flagships.

Specs

• Context length: 1,000,000 tokens
• Deep reasoning, with support for structured output
• Model ID: qwen3.8-max (compatibility alias: qwen/qwen3.8-max)

How to call it

curl https://bazaarlink.ai/api/v1/chat/completions \
  -H "Authorization: Bearer sk-bl-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.8-max",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

OpenAI SDK compatible; just change base_url and model, no other code changes needed.

When to choose Qwen3.8-Max, and when not to

Good fit:

• Deep reasoning, math, and code analysis tasks where cost is the main consideration
• Cases where you have already evaluated Kimi K3 / GLM-5.3 but found that the extra cost does not buy a proportional quality gain

Not a good fit:

• Latency-sensitive real-time scenarios: at 39.0 tokens/second, it is the slowest of the three Chinese-camp flagships
• Tasks that require visual input: the specs compared in this article cover only text, so for multimodal needs, first confirm the actual supported scope

Sources and query log

• Artificial Analysis: Qwen3.8 Max model page (query date 2026-09-20)
• BazaarLink Models: Qwen3.8-Max (unit prices and specs; query date 2026-09-02)
• Further reading: How do I expense AI API costs in Taiwan? 2026 complete guide, DeepSeek API TWD cost estimate 2026

TWD billing · Taiwan invoices · leading AI models · OpenAI-compatible API