Model card
Niah 1 Fast
niah-1-fast
Best frontier / speed mix — latency-first with full tools.
API
- model
niah-1-fast- endpoint
-
POST https://api.niah.si/v1/chat/completions
curl -s https://api.niah.si/v1/chat/completions \
-H "Authorization: Bearer $NIAH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"niah-1-fast","messages":[{"role":"user","content":"Hello"}]}'
Intended use
Latency-sensitive turns — titles, quick answers, and high-volume loops at the best frontier / speed balance.
How it works
Niah's best frontier / speed mix: full tools with a tighter turn budget, optimized for latency. Backends rotate toward whatever currently delivers the best speed / quality tradeoff.
Positioning: Best frontier / speed mix
Capabilities
- Streaming
- OpenAI-compatible
- Tools
- Web search
- Image generation
- Multi-turn agents
Evaluations
Scores appear here when published. Until then, status stays planned.
-
LMArena (Chatbot Arena) planned
Submit niah-1-fast as a public API black box when the SKU is frozen for eval.
-
Artificial Analysis / academic suite planned
Run the shipped SKU as a product black box (prompts + tools + orchestration) vs naked frontier models — we rank the product, not stripped base weights.
Pricing
Billed as Niah API / plan AI usage (not a separate provider invoice). See niah.si/pricing for plans and prepaid credits.
View pricingLimits & safety
Not a base LLM. Responses inherit provider safety systems for the active frontier plus Niah org policy, rate limits, and usage controls. Do not treat as a downloadable weight release.
Changelog
- — Public SKU cards published (tier positioning; underlying frontiers may rotate).
Other SKUs
We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track