Novita AI
Primary
Total 88/100 · Cost 85 · Avail 100 · Integrity 100
主路：；价格处于同模型低位；近期可用性证据较好；协议/检测证据较充分。
Input $0.5100/M, Output $0.7400/M
llama-3-70b-instruct Price Comparison
Input Price Range
$0.4000 ~ $0.9000/M
Output Price Range
$0.5600 ~ $0.9000/M
Input Price Avg
$0.6033/M
Price Spread
2.2x
AI API Routing Candidates
Based on pricing, protocol evidence, availability, model coverage, and risk signals, the candidate providers for llama-3-70b-instruct are classified as primary, fallback, watch, or avoid. This page provides routing intelligence only — we do not host API keys or forward real requests.
Fireworks AI
Fallback
Total 74/100 · Cost 30 · Avail 100 · Integrity 100
备路：；价格优势不明显；近期可用性证据较好；协议/检测证据较充分。
Input $0.9000/M, Output $0.9000/M
RunAPI:高质量API中转平台
Fallback
Total 61/100 · Cost 100 · Avail 35 · Integrity 55
备路：；价格处于同模型低位；风险标记：low_detection_sample。
Input $0.4000/M, Output $0.5600/M
Test endpoint:
/api/router/candidates?model=llama-3-70b-instruct. Fields include cost / integrity / availability / coverage / risk evidence.
Turn candidates into executable config:
Router Config Lab →
Generate LiteLLM / OpenRouter / Cloudflare AI Gateway / Portkey routing drafts with env placeholders — no keys stored.
📊 Price Comparison Across Relays
| Relay | Raw Model ID | Input /M | Output /M | Cache /M | Source / Time | Details |
|---|---|---|---|---|---|---|
| RunAPI:高质量API中转平台 | meta-llama/llama-3-70b-instruct |
$0.4000 | $0.5600 | — |
https://runapi.co/api/pricing |
View → |
| Novita AI | meta-llama/llama-3-70b-instruct |
$0.5100 | $0.7400 | — |
litellm |
View → |
| Fireworks AI | accounts/fireworks/models/llama-v3-70b-instruct |
$0.9000 | $0.9000 | — |
litellm |
View → |
🧬 Other SKUs in this product line (1)
Other independently-billable SKUs under the llama-3-70b product line (e.g. chat / instruct / different context tiers). Pricing and available relays vary.
| SKU | Form | Priced Relays | Min Input /M | Details |
|---|---|---|---|---|
llama-3-70b |
standard | 12 | $0.0000 | View prices → |
🔗 Aliases for this model (1)
Different relays use different names for the same model. The aliases below all point to llama-3-70b-instruct. Visiting /en/prices/{alias} redirects here.
llama-v3-70b-instruct
📈 Price History Trend (30 days)
Window 2026-06-30 → 2026-07-30 · 23 sample days · Price stable (within ±1%)
Source: price_history_snapshot · daily cron tools_dev/cron_scripts/price_snapshot_daily.py · cross-relay aggregated input price distribution