Your context fits. Your route should too.

Long-context work breaks when the provider boundary is unclear. AIWave keeps the client familiar and makes the route, token classes, cache state and charge basis readable.

Run a real workload Check the V4 Pro route

Three things to know before the first large request.

Context size is only one constraint. The request must also survive the route, the cache prefix, the output limit, and the account ledger.

READ THE FAILURE BEFORE YOU SCALE

ContextModel fact

Pin the tested model ID.

Record the context and output limits with the date of your acceptance test. An alias is not a guarantee.

CachePrefix reuse

Keep repeated context stable.

Cache hit and miss fields explain why two similar requests can have different cost and first-token behavior.

Recovery429 · overflow

Make the failure path explicit.

Trim, retry, back off or switch the model only when the application has a rule for that state.

Cost starts with the input you actually send.

Reference rates for the two long-context routes are checked 2026-08-27. Use the worksheet below for a first estimate, then reconcile the result against the request ledger.

OFFICIAL DEEPSEEK REFERENCE · 2026-08-27

Official DeepSeek reference rates for two long-context routes
Route	Input / 1M	Cache hit / 1M	Output / 1M
deepseek-v4-flash	$0.638	$0.0202884	$1.914
deepseek-v4-pro	$1.914	$0.0637362	$5.742

Context Cost Worksheet

Static v1 worksheet. Fill in your own workload values; this page does not submit or store them.

Input tokensCache-hit tokensOutput tokensRequests per month

Worksheet formula: input tokens × input rate + cache-hit tokens × cache-hit rate + output tokens × output rate, multiplied by requests. Confirm the selected model, group and live rate before funding.

Keep your client. Change the route.

Set the base URL, preserve the OpenAI request shape, then add a workload-specific overflow and 429 policy.

MIGRATION ROUTE

Python · first request

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://aiwave.live/v1",
    api_key=os.environ["AIWAVE_API_KEY"],
)

response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[{"role": "user", "content": "Summarize the repository context."}],
    max_tokens=2048,
)
print(response.usage)

Before production.
Test the maximum context, cache behavior, timeout, retry policy and ledger output with the same payload shape your application will send.

Start with V4 Pro. Switch models in one line — no new account, same USD invoice.

AIWave is the route around the workload, not a commitment to one model family.

Open the V4 Pro route Read the rate card