Qwen: Qwen3 Next 80B A3B Thinking

Text input Text output
Author's Description

Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model in the Qwen3-Next line that outputs structured “thinking” traces by default. It’s designed for hard multi-step problems; math proofs, code synthesis/debugging, logic, and agentic planning, and reports strong results across knowledge, reasoning, coding, alignment, and multilingual evaluations. Compared with prior Qwen3 variants, it emphasizes stability under long chains of thought and efficient scaling during inference, and it is tuned to follow complex instructions while reducing repetitive or off-task behavior. The model is suitable for agent frameworks and tool use (function calling), retrieval-heavy workflows, and standardized benchmarking where step-by-step solutions are required. It supports long, detailed completions and leverages throughput-oriented techniques (e.g., multi-token prediction) for faster generation. Note that it operates in thinking-only mode.

Key Specifications
Cost
$$$$$
Context
262K
Parameters
80B
Released
Sep 11, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Reasoning Response Format Presence Penalty Seed Temperature Top P Max Tokens Tool Choice Tools Include Reasoning
Features

This model supports the following features:

Response Format Tools Reasoning
Performance Summary

Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model designed for complex multi-step problems, emphasizing structured "thinking" traces. While it demonstrates exceptional reliability with a 99% success rate, its speed performance is notably slower, ranking in the 12th percentile, and it is positioned at premium pricing levels (5th percentile). The model exhibits strong performance in specialized areas. It achieves high accuracy in Coding (94.9%, 93rd percentile), Reasoning (96.0%, 86th percentile), and General Knowledge (99.5%, 72nd percentile). Its Ethics performance is perfect at 100% accuracy, making it the most accurate model at its price point and among models of similar speed. Hallucinations are well-managed with 98.0% accuracy. However, a significant weakness is its Instruction Following capability, where it scores only 14.7% accuracy (25th percentile), indicating challenges with complex, multi-layered instructions. Mathematics performance is moderate at 88.9% (61st percentile). This model is best suited for applications requiring detailed, step-by-step solutions and high reliability, particularly in agent frameworks and tool use, where its "thinking-only" mode can be leveraged despite its slower response times and higher cost.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.5
Completion $6

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Alibaba
Alibaba | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.5 / 1M tokens $6 / 1M tokens
Novita
Novita | qwen/qwen3-next-80b-a3b-thinking-2509 131K $0.15 / 1M tokens $1.2 / 1M tokens
Chutes
Chutes | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
DeepInfra
DeepInfra | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
Hyperbolic
Hyperbolic | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.3 / 1M tokens $0.3 / 1M tokens
GMICloud
GMICloud | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
AtlasCloud
AtlasCloud | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.5 / 1M tokens
NCompass
NCompass | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
Together
Together | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.5 / 1M tokens
Parasail
Parasail | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
Google
Google | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
Parasail
Parasail | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
Chutes
Chutes | qwen/qwen3-next-80b-a3b-thinking-2509 262K $0.15 / 1M tokens $1.2 / 1M tokens
Novita
Novita | qwen/qwen3-next-80b-a3b-thinking-2509 131K $0.15 / 1M tokens $1.5 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by qwen