Qwen3-Next-80B-A3B-Instruct: Benchmarks, Pricing & Context Window
Qwen3-Next-80B-A3B-Instruct is a language model from Qwen, released in September 2025, with a 262K-token context window, and pricing from $0.090/M input and $1.10/M output.
Qwen3-Next-80B-A3B-Instruct is the first in the Qwen3-Next series, featuring groundbreaking architectural innovations. It uses Hybrid Attention combining Gated DeltaNet and Gated Attention for efficient ultra-long context modeling,
Qwen3-Next-80B-A3B-Instruct benchmarks
Capability tiers
Standing within each category, adjusted for leaderboard depth.
Real tasks performance
High-confidence performance for Qwen3-Next-80B-A3B-Instruct across real-world prompt categories. Only 95% intervals at most 4 points wide are shown.
Performance by conversation depth
How Qwen3-Next-80B-A3B-Instruct holds up as conversations get longer.
Quality Tracker
Qwen3-Next-80B-A3B-Instruct Performance Across Datasets
Scores sourced from the model's scorecard, paper, or official blog posts
llm-stats.com - Fri Oct 02 2026Qwen3-Next-80B-A3B-Instruct pricing
Providers
Qwen3-Next-80B-A3B-Instruct starts at $0.0900 per million input tokens and $1.10 per million output tokens via DeepInfra. See all 2 providers below with their per-token pricing, latency, throughput, and modality support.
| Provider | Input $/M | Cached input $/M | Output $/M | Context in / out | TTFT p95 s | Output p5 c/s | Modalities in / out |
|---|---|---|---|---|---|---|---|
 DeepInfra | $0.0900 | — | $1.10 | 262.1K/262.1K | — | — | / |
 Novita | $0.150 | — | $1.50 | 65.5K/65.5K | — | — | / |
Cached input is the discounted price for prompt tokens served from a provider cache. TTFT is time to first token. Output is characters per second; p5 is the sustained floor exceeded by 95% of observed requests.
Qwen3-Next-80B-A3B-Instruct model size
Qwen3-Next-80B-A3B-Instruct has 80 billion parameters and was trained on 15 trillion tokens. See how it compares to other models in the same parameter range.
Qwen3-Next-80B-A3B-Instruct context window
Input and output token limits for Qwen3-Next-80B-A3B-Instruct, plus how it ranks on long-context understanding.
Try now
huggleMake it with
Qwen3-Next-80B-A3B-Instruct.
Qwen3-Next-80B-A3B-Instruct latency
Qwen3-Next-80B-A3B-Instruct time to first token, sustained output throughput, and failed-request rate from live model usage over the trailing 7 days.
Qwen3-Next-80B-A3B-Instruct examples
Recent arena outputs from Qwen3-Next-80B-A3B-Instruct, picked from the highest-ranked matchups.
Qwen3-Next-80B-A3B-Instruct license
Qwen3-Next-80B-A3B-Instruct is released under the Apache 2.0 license, which permits commercial use, has 80.0B parameters.
- License
- Apache 2.0
- Commercial use allowed
- Parameters
- 80.0B
Apache License 2.0 - allows commercial use
Qwen3-Next-80B-A3B-Instruct resources
Official sources for Qwen3-Next-80B-A3B-Instruct: provider documentation, official playground, official launch post, source repository.
Qwen3-Next-80B-A3B-Instruct vs other models
The most-compared alternatives to Qwen3-Next-80B-A3B-Instruct are Qwen3 14B, Llama 3.1 70B Instruct, Qwen3 VL 32B Thinking. Open any pair side-by-side for benchmarks, pricing, context, and latency.
Models like Qwen3-Next-80B-A3B-Instruct
Models ranked just above and below Qwen3-Next-80B-A3B-Instruct by LLM Stats score.
FAQ
Common questions about Qwen3-Next-80B-A3B-Instruct.