StarSeaAPI
claude-sonnet-5 vs qwen3.8-max
| claude-sonnet-5 | qwen3.8-max | |
|---|---|---|
| Providers | Anthropic | Alibaba Qwen |
| Input $/M | $2.00 | $1.82 |
| Output $/M | $10.00 | $5.45 |
| Output multiplier | ×5.00 | ×3.00 |
| Context window | 200K tokens | 1.05M tokens |
| Intelligence index | 45.1 | 53.4 |
| Coding index | 71.5 | 68.9 |
What the numbers mean
Input pricing is close — 1.10× apart — so the decision should come down to measured quality on your own evaluation set rather than headline cost.
On measured intelligence index the gap is 8.3 points, with qwen3.8-max ahead. That is a large enough spread to be visible on multi-step tasks.
Tags: claude-sonnet-5 — Chat, Reasoning, Coding, Vision · qwen3.8-max — Chat, Reasoning, Coding, Vision
Code
from openai import OpenAI
client = OpenAI(base_url="https://starseaapi.com/v1", api_key="sk-...")
# claude-sonnet-5
client.chat.completions.create(model="claude-sonnet-5", messages=[{"role":"user","content":"Hi"}])
# qwen3.8-max
client.chat.completions.create(model="qwen3.8-max", messages=[{"role": "user", "content": "Hi"}])