Qwen3 Next 80B A3B Instruct is an open-weight, 80-billion-parameter mixture-of-experts model that activates only about 3 billion parameters per token, sharply reducing compute while keeping strong performance. It uses a hybrid attention design combining Gated DeltaNet and Gated Attention, with 512 experts and 10 activated per token, and reaches results comparable to much larger models on knowledge, reasoning, and coding benchmarks. The model runs in instruct mode only and does not emit reasoning traces, so it favors rapid, direct responses. Alibaba reports more than 10x higher inference throughput than Qwen3-32B for contexts beyond 32K tokens, making it efficient on long inputs. With a 131K context window, it is well suited to production RAG, tool-calling agents, and agentic orchestration where cost efficiency and latency matter. Its open-weight release allows local deployment and fine-tuning for specialized domains.
Qwen3 Next 80B A3B Instruct runs on 2 different routes through the Opper gateway. Compare residency, ZDR, and training posture at a glance β full data-handling detail per route below.
Training posture across routes: No training on prompts by default.
Each route hosting Qwen3 Next 80B A3B Instruct has its own privacy posture, residency, and GDPR terms. Postures are maintained by Opper with a last-verification timestamp.
Zero data retention is available via Opper Enterprise contract. No training on customer data. US; unknown.
Zero data retention is available via Opper Enterprise contract. No training on customer data. US; SCCs; DPA available.
Independent benchmark scores β composite indices for reasoning, coding, and math, plus individual eval scores where available.
Call Qwen3 Next 80B A3B Instruct through the Opper gateway with one API key. Let your coding agent set it up, or call it directly β Opper is drop-in compatible with the OpenAI, Anthropic, and Google AI SDKs.
Copy this and paste it into your coding agent β Claude Code, Cursor, Codex, and more β and it'll wire up Opper for you.
import OpenAI from "openai";const client = new OpenAI({apiKey: process.env.OPPER_API_KEY,baseURL: "https://api.opper.click/v3/compat",});const completion = await client.chat.completions.create({model: "deepinfra/Qwen/Qwen3-Next-80B-A3B-Instruct",messages: [{ role: "user", content: "Hello" }],});console.log(completion.choices[0].message.content);
Side-by-side on privacy, EU hosting, pricing, and benchmarks.
