Fikra Embeddings is live in the Fikra catalogue as qwen-embedding: $0.05 input and $0.0 output per 1M tokens. Best for: retrieval, semantic search and RAG. Route it with one Fikra key at https://api.fikraapi.co.ke/v1.
Fikra Embeddings is live in the Fikra catalogue as qwen-embedding: $0.05 input and $0.0 output per 1M tokens. Best for: retrieval, semantic search and RAG. Route it with one Fikra key at https://api.fikraapi.co.ke/v1.
Pricing and limits (live registry)
KES figures use KES 133 = $1, the published checkout rate. Your card issuer sets its own conversion rate. Prices change with provider terms, check /models before a production commitment.
Field
Value
KES at checkout rate
Input / 1M tokens
$0.05
KES 6.65
Cached input / 1M tokens
not listed
-
Output / 1M tokens
$0.0
KES 0.00
Context window
not listed
See /models
Max output
not published
See /models
Rate limit
500 req/min
50000 req/day
What it costs on your workload
10M input tokens and 2M output tokens costs roughly $0.50 at the listed rates (KES 66.50 at KES 133 = $1), before any cache discount.
Output tokens cost more than input on most routes. Cap max_tokens to what the answer actually needs.
Cached input is cheaper when your prompt has a stable system-level prefix.
Score cost per accepted task, not cost per token. A route that fails one task in five is not cheaper at any sticker price.
How to run it
Set base_url to https://api.fikraapi.co.ke/v1 in any OpenAI-compatible client and use model=qwen-embedding. The key is from your dashboard; it is shown once.
Confirm qwen-embedding is listed at /v1/models for your account before committing traffic.
Start with a short max_tokens and temperature 0 for classification and extraction.
Keep the request ID from the response if you need to reconcile a wallet charge.
Before production: run at least 50 of your own prompts and log cost per accepted task.
curl https://api.fikraapi.co.ke/v1/chat/completions \
-H "Authorization: Bearer YOUR_FIKRA_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"qwen-embedding","messages":[{"role":"user","content":"Summarise this ticket in two lines."}]}'
How this route compares
For text work, compare accepted-task cost against deepseek-v4.1-flash-1017 ($0.05 input / $0.15 output per 1M) and glm-5.3-flash ($0.1875 input / $0.625 output). For hard reasoning, compare against deepseek-v4-pro-pinned. For retrieval, pair any chat route with qwen-embedding.
Freeze a 50-prompt set from your real traffic, including 10 failure-shaped prompts.
Log model ID, request ID, tokens, cache hits and total cost for every call.
Keep the route with the lowest wallet spend per accepted task. Re-run when /models lists a revision.
Before you commit budget
Plans are prepaid per 30 days: Lite $10, Pro $25, Max $60. The balance is the hard cap; calls stop when funds run out rather than overcharge. Enterprise teams with volume or contract requirements can contact support before migrating production traffic.
Common questions
Is qwen-embedding available in Kenya?
Availability is per route, not per country. Check /v1/models; if the route is listed, the same key works everywhere. Kenya buyers fund with M-Pesa, Airtel Money or card; other markets pay by card.
How is this route billed?
Per the live registry: USD per 1M input and output tokens when billed per token, with cached input priced separately where listed. Minimum top-up is $5; the wallet credits in USD after checkout.
What if the route disappears from /v1/models?
Stop sending production traffic to it. Routes are withdrawn with capacity or licensing changes. Contact support with the request ID and status code.