Qwen3 Next 80B A3B Instruct
Qwen · released Sep 9, 2025 · Qwen/Qwen3-Next-80B-A3B-Instruct
- Type
- Open weightsApache License 2.0
- Params
- 81.3B
- Context
- 262K
3B active per word · about 197K words of context
Our take
Written Sep 2, 2026Qwen3 Next is a text-only downloadable model from Alibaba with a permissive Apache licence. It uses a mixture-of-experts design where only 3 billion of its 81.3 billion total parameters activate per token, making it extremely parameter-efficient. Its measured coding and hard-prompt performance sits well above its overall arena baseline, though creative writing is a clear weak spot.
Pick this for Apache-licensed self-hosting or budget hosted inference where you want MoE efficiency rather than peak arena rank. Use it for coding, maths or long-context text work up to 262,144 tokens. Skip it if creative writing quality matters, if you need multimodal input, or if you want measured throughput on every host.
The case for it
- Extreme parameter efficiency: only 3 billion active per token from 81.3 billion total, roughly 27× sparsity.
- Permissive Apache 2.0 licence allows commercial use, fine-tuning and redistribution.
- Coding performance leads its own overall score by a notable margin on the arena leaderboard.
- Hard-prompt and maths scores also sit measurably above its overall baseline.
The case against it
- Creative writing is its clear weak spot, sitting 86 points below its own overall arena score.
- Throughput varies widely where known, and three of nine tracked offers have no measured data.
- Output pricing spans more than a third between cheapest and most expensive host for identical weights.
How good is it?
EverydayGeneral questions and everyday reasoning
Arena Text (overall)97th of 168 · 1399
CodingWriting and fixing code on its own
Arena Coding98th of 168 · 1444
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent.
WritingDrafting and rewriting prose
Arena Creative Writing121st of 168 · 1313
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model6 scoresEvery figure we hold, from 6 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Comfortable fit
Apple M2 Max (38-core GPU) · 96 GB
Room to spare. 17.8 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked between 4 hours and 3 days ago — each listing carries its own date.
Cheapest of the 2 listings we can compare like for like — at 262K of context, out of 6 in the table below. One cheaper row there is outside that comparison: a different context length.
- per 1M tokens
- $0.10 in / $1.10 out
- Context served
- 262K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| Alibaba CloudThrough OpenRouter | $0.098 / $0.78checked 4 hours ago | 131K33K max reply | 50 tok/s | No | Yesunknown period | Unknown |
| DeepInfrafp8Direct and through OpenRouter | $0.090 / $1.10checked 4 hours ago | 262K16K max reply through OpenRouter | 43 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| OpenRouterOpenRouter's own listing | $0.10 / $1.10checked 4 hours ago | 262K | not measured | Unknown | Unknown | Unknown |
| Parasailfp8Through OpenRouter | $0.10 / $1.10checked 4 hours ago | 262K236K max reply | 4 tok/s | No | No | Confirmed |
| Google Vertex AIglobalThrough OpenRouter | $0.15 / $1.20checked 4 hours ago | 262K236K max reply | 10 tok/s | No | No | Confirmed |
| Novita AIbf16Direct and through OpenRouter | $0.15 / $1.50checked 4 hours ago directchecked 3 days ago through OpenRouter | 131K33K max reply through OpenRouter | 3 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
DeepInfra
Google Vertex AI
Novita AIAcross the 6 listings we hold: 5 say they do not train on prompts (2 of them only through OpenRouter), 0 say they do and 1 does not say. 4 appear in the zero-retention registry we check (2 of them only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| Alibaba CloudThrough OpenRouter | ✓ | ✓ | ✗ |
| DeepInfrafp8Direct and through OpenRouter | ✓ | ✓ | ✓ |
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| Parasailfp8Through OpenRouter | ✓ | ✓ | ✓ |
| Google Vertex AIglobalThrough OpenRouter | ✓ | ✓ | ✓ |
| Novita AIbf16Direct and through OpenRouter | ✓ | ✓ | ✗ |
Tool calling: 6 of 6 listings say yes. JSON output: 6 of 6 listings say yes. Strict schema: 4 of 6 listings say yes, 2 say no.
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 6 listings does not say whether it trains on prompts, and 2 answer only through OpenRouter, not for their own listing.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- Qwen/Qwen3-Next-80B-A3B-Instruct
- Architecture
- Mixture of experts
- Takes in, gives back
- Text in, text out
- Catalogue slug
- qwen-qwen3-next-80b-a3b-instruct