1,000 Qwen3.8 27B (OpenRouter) tokens cost $0.00042 as input or $0.003 as output. Most workloads land in between — a 70/30 input/output split costs about $0.001194. Verified Aug 28, 2026
1,000 tokens is the rough size of a short article, a few code snippets, or a handful of chat turns — about 0.2 chat-style requests (4,000 input + 800 output tokens each). Billed on Qwen3.8 27B (OpenRouter) at $0.42/M input and $3.00/M output, that volume costs $0.00042 as pure input, $0.003 as pure output, or about $0.001194 at a typical 70/30 mix. Qwen3.8 27B (OpenRouter) has no published cached-input discount, so repeated context bills at the full rate. As a balanced-tier model it is a common production default for exactly this kind of volume.
Interactive calculator
1,000 in · 0 out
Input price
$0.42 / 1M tokens
Output price
$3.00 / 1M tokens
Uncached pricing. Add caching and monthly projections in the full calculator.
What 1,000 tokens buys in practice
$0.000936
80% input / 20% output — short replies over conversation history.
$0.000678
90% input / 10% output — long retrieved context, grounded answers.
$0.000549
95% input / 5% output — full documents in, structured extraction out.
Same 1,000 tokens on other models
Related calculations
FAQ
1,000 input tokens cost $0.00042; 1,000 output tokens cost $0.003. A typical 70/30 input/output mix costs about $0.001194. Rates: $0.42/M input, $3.00/M output.
A chat-style request of ~4,800 tokens (4,000 in + 800 out) means 1,000 tokens is roughly 0.2 requests. For a RAG workload with ~8,500 tokens per request, about 0.1 requests.
Qwen3.8 27B (OpenRouter) has no published cached-input discount on its official pricing page; all input is billed at $0.42/M.