Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage & billing
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Usage & billing
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksNano RPCEarn crypto
  1. Models
  2. Text
  3. Qwen

Qwen3 Next 80B A3B (Instruct)

Based on the new Qwen3‑Next architecture (hybrid attention, highly sparse MoE, training‑stability optimizations, and multi‑token prediction), the Qwen3‑Next‑80B‑A3B‑Instruct model delivers extreme efficiency with only 3B active parameters per pass. It performs comparably to Qwen3‑235B‑A22B‑Instruct‑2507 and shows clear advantages on ultra‑long context tasks (up to 256K tokens).

  • Tool Calling
  • Structured Output

Added Sep 11, 2025

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.09
Output
$1.10
Compare provider prices

Specifications

Context window
256K
Max output
262.1K
Parameters
80B / 3B
Total / active
Avg output (7d)
845 tokens
Longer than 63% of models

Benchmarks

Sourced from LMArena.

Arena Score

1398.7

Overall Rank

#169 / 421

Votes

22,692

Confidence Interval

1393.9 - 1403.5

Category Scores

Coding

#168 / 416

4,802 votes

1444.3

Math

#141 / 404

1,199 votes

1413.3

Longer Query

#188 / 399

5,372 votes

1389.6

Creative Writing

#224 / 419

3,018 votes

1313.2

Instruction Following

#185 / 421

6,390 votes

1376.4

Hard Prompts

#165 / 421

11,956 votes

1419.5

Additional Categories
22

Japanese

#124 / 288

318 votes

1369.3

Industry Medicine And Healthcare

#131 / 390

1,179 votes

1440.2

Spanish

#138 / 303

684 votes

1408.4

Industry Business And Management And Financial Operations

#145 / 414

4,276 votes

1412.9

German

#148 / 320

490 votes

1387.3

Chinese

#151 / 395

1,719 votes

1437.6

Industry Mathematical

#155 / 401

1,062 votes

1411.8

Industry Software And It Services

#159 / 421

8,030 votes

1441.2

Industry Legal And Government

#161 / 393

1,412 votes

1408.7

Industry Life And Physical And Social Science

#165 / 419

3,519 votes

1417.0

Non English

#167 / 421

12,946 votes

1384.9

Hard Prompts English

#168 / 419

5,204 votes

1426.7

Polish

#168 / 244

1,025 votes

1369.5

Korean

#169 / 289

520 votes

1325.4

Multi Turn

#169 / 419

4,006 votes

1400.2

Exclude Ties

#170 / 421

16,005 votes

1381.1

Russian

#170 / 385

1,605 votes

1388.2

French

#171 / 301

384 votes

1401.9

English

#177 / 421

9,712 votes

1407.6

Expert

#185 / 371

1,084 votes

1400.7

Industry Writing And Literature And Language

#208 / 420

5,040 votes

1342.2

Industry Entertainment And Sports And Media

#212 / 419

4,203 votes

1325.1

Published 2026-11-09 · Matched as Qwen3-Next-80B-A3B-Instruct

LMArena Dataset

Providers

Choose or restrict providers with the API. Provider-selection prices include the selection markup. Provider selection guide · Pricing API (JSON, USD per 1K tokens)

Auto
Auto (ZDR)
$0.09$1.10N/AN/AN/A
Auto
Auto (Anonymized)
$0.09$1.10N/AN/AN/A
DeepInfra
DeepInfra
FP8
US
$0.0945$1.16N/A0.5s109 tps
Parasail
Parasail
FP8
US
$0.105$1.16$0.07350.4s95 tps
Privacy:ZDR, prompts not retainedNo training on promptsBest latency and throughput are highlighted.
Batch API
ProviderInputOutputCached input
Auto
Autoqwen/qwen3-next-80b-a3b-instruct
$0.09/M$1.10/MN/A

Batch rows for this model use normal routing and cost the same as standard API requests. Set provider in a row to choose a provider.

Related text models

Compare Qwen3 Next 80B A3B (Instruct) with similar models from the same provider or model family.

Qwen3 Next 80B A3B (Thinking)

qwen/qwen3-next-80b-a3b-thinking

Qwen3 Next 80B A3B (Thinking)

Qwen3.6 35B A3B Thinking

qwen/qwen3.6-35b-a3b:thinking

Qwen3.6 35B A3B is a native vision-language MoE model with hybrid attention. Compared to Qwen3.5 35B A3B, Alibaba reports stronger agentic coding, mathematical and code reasoning, and better spatial understanding (including object localization and detection).

Qwen3.6 35B A3B

qwen/qwen3.6-35b-a3b

Qwen3.6 35B A3B is a native vision-language MoE model with hybrid attention. Compared to Qwen3.5 35B A3B, Alibaba reports stronger agentic coding, mathematical and code reasoning, and better spatial understanding (including object localization and detection).

Qwen3 Coder Next

qwen/qwen3-coder-next

Qwen3 Coder Next is an open-weight coding model built on Qwen3-Next-80B-A3B-Base (hybrid attention + MoE). It is agentically trained at scale on executable tasks and environment interaction, delivering strong coding and tool-use performance at lower inference cost. Native 256K context.

Qwen 3 235b A22B 2507

qwen/qwen3-235b-a22b-instruct-2507

Qwen 3 235b A22B Instruct 2507 the updated version of Qwen3 235B A22B, with significant improvements in performance. This model is non-thinking.

Qwen3 30B A3B

qwen/qwen3-30b-a3b

Qwen 3 30b A3B is a 30b model with 3 billion active parameters per pass. Supports switching between thinking and non thinking: trigger thinking with /think and /no_think anywhere in a prompt or system message to toggle chain-of-thought reasoning.