Alibaba: qwen3-next-80b-a3b-instruct:free
The flagship instruction-tuned model of the Qwen3-Next series, released by Alibaba Cloud on September 11, 2025. It features a revolutionary Hybrid Attention architecture, combining Gated DeltaNet (linear attention) and standard Gated Attention. With 80B total parameters, it activates only 3.9B per token during inference (A3B). Designed for Non-Thinking tasks, this model skips visible reasoning traces in favor of near-instant, high-precision answers. Trained on 15 trillion tokens, its performance rivals the Qwen3-235B flagship while offering superior efficiency for ultra-long context (256k+) and high-concurrency Agentic workflows. It stands as the premier price-performance leader for production-grade AI deployments in 2026.
Prezzi
Performance
Utilizzo e classifica
Parametri supportati
Tutti i provider = supportato da ogni upstream che serve questo modello. Alcuni provider = dipende dall'upstream che gestisce la richiesta. Predefinito = il valore inviato quando non lo imposti.
| Parametro | Provider | Predefinito |
|---|---|---|
| frequency_penalty | Alcuni provider | - |
| logit_bias | Alcuni provider | - |
| logprobs | Alcuni provider | - |
| max_tokens | Tutti i provider | - |
| min_p | Alcuni provider | - |
| presence_penalty | Tutti i provider | - |
| repetition_penalty | Alcuni provider | - |
| response_format | Tutti i provider | - |
| seed | Tutti i provider | - |
| stop | Alcuni provider | - |
| structured_outputs | Alcuni provider | - |
| temperature | Tutti i provider | - |
| tool_choice | Tutti i provider | - |
| tools | Tutti i provider | - |
| top_k | Alcuni provider | - |
| top_logprobs | Alcuni provider | - |
| top_p | Tutti i provider | - |