TackleKey
Cost-aware routing

Cheapest successful route candidates

A public, safe ranking of lower-cost candidates from TackleKey pricing data. The actual best route still depends on success, cooldown, endpoint support, and request context.

Public low-cost candidates

#ModelProviderInput priceBoundary
1mistralai/mistral-nemoMistral$0.02 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
2deepseek/deepseek-ocrDeepSeek$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
3baidu/ernie-4.5-0.3b百度$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
4meta-llama/llama-3.1-8b-instructMeta$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
5minimax/minimax-m2MiniMax$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
6minimax/minimax-m2.1MiniMax$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
7minimax/minimax-m2.5MiniMax$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
8minimax/minimax-m2.7MiniMax$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
9minimax/minimax-m3MiniMax$0.04 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
10qwen/qwen-2.5-7b-instruct阿里巴巴$0.06 / 1M tokensPublic pricing candidate; confirm live route behavior before scale.
Private source names, account pools, tokens, and raw internal failures are not published. Future versions can add sanitized aggregate success-rate data.

Try TackleKey with one OpenAI-compatible change

Create an account, generate a project API key, then replace your client base URL with the TackleKey endpoint. Keep keys server-side and verify live pricing before scaling traffic.