llama-3.3-70b

chat Open weights id: llama-3.3-70b

Route	Residency	₹ in / Mtok	₹ out / Mtok	Live	p50 / p95	tokens/s	7d success
Groq	global	₹57	₹76	…	—	—	—
OpenRouter BYOK	global	₹10	₹31	…	—	—	—
Fireworks AI BYOK	global	₹86	₹86	…	—	—	—

Latency, throughput and success rate are measured from real traffic over the trailing 7 days (/v1/models/llama-3.3-70b/stats) — compare providers, then chain them: the failover playground proves this model keeps answering when its first provider goes down.

curl https://api.bharatrouter.com/v1/chat/completions \
  -H "Authorization: Bearer br-..." -H "Content-Type: application/json" \
  -d '{"model": "llama-3.3-70b", "messages": [{"role": "user", "content": "namaste"}], "data_policy": "india_only"}'

llama-3.3-70b is served on BharatRouter through 3 routes with automatic failover — same OpenAI-compatible API, INR billing, and optimize routing (price / latency / quality) as every other model in the catalog.

Open-weight model — licence and weights on Hugging Face ↗

Card maintained by BharatRouter. Languages: en, hi, multilingual.