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

Qwen3.5 35B A3B

Qwen3.5 35B A3B is a native vision-language MoE model with hybrid attention designed for efficient inference and strong general performance.

  • Vision
  • Video Input

Added Feb 24, 2026

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.14
Output
$1.00
Cache read
$0.05
Compare provider prices

Specifications

Context window
260.1K
Max output
65.5K
Parameters
35B / 3B
Total / active
Avg output (7d)
211 tokens
Longer than 15% of models

Benchmarks

Sourced from LMArena.

Arena Score

1394.0

Overall Rank

#175 / 421

Votes

30,206

Confidence Interval

1389.7 - 1398.2

Category Scores

Coding

#182 / 416

8,314 votes

1434.6

Math

#169 / 404

1,835 votes

1396.9

Longer Query

#180 / 399

11,776 votes

1399.4

Creative Writing

#194 / 419

4,617 votes

1342.4

Instruction Following

#168 / 421

9,838 votes

1386.9

Hard Prompts

#180 / 421

19,164 votes

1412.3

Additional Categories
22

Chinese

#142 / 395

2,064 votes

1447.8

French

#147 / 301

873 votes

1422.4

Korean

#151 / 289

485 votes

1346.0

Industry Medicine And Healthcare

#158 / 390

2,254 votes

1427.0

Expert

#161 / 371

2,775 votes

1420.9

Polish

#162 / 244

584 votes

1373.7

Japanese

#166 / 288

248 votes

1312.7

Industry Business And Management And Financial Operations

#167 / 414

6,084 votes

1402.7

Industry Mathematical

#167 / 401

1,708 votes

1409.7

Spanish

#167 / 303

992 votes

1387.1

Industry Life And Physical And Social Science

#169 / 419

5,157 votes

1416.0

Hard Prompts English

#170 / 419

8,757 votes

1425.5

Industry Software And It Services

#172 / 421

11,950 votes

1429.9

English

#173 / 421

13,445 votes

1410.3

Exclude Ties

#173 / 421

22,025 votes

1373.7

German

#174 / 320

484 votes

1360.3

Industry Legal And Government

#175 / 393

2,361 votes

1398.8

Multi Turn

#177 / 419

5,450 votes

1394.3

Non English

#178 / 421

16,761 votes

1374.2

Russian

#185 / 385

3,612 votes

1375.5

Industry Writing And Literature And Language

#186 / 420

6,995 votes

1364.4

Industry Entertainment And Sports And Media

#191 / 419

5,884 votes

1344.9

Published 2026-11-09 · Matched as Qwen3.5-35B-A3B

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)
$0.14$1.00$0.05N/AN/A
Auto
Auto (Anonymized)
$0.14$1.00$0.05N/AN/A
Darkbloom
Darkbloom
TEE attested
FP4
GLB
$0.0682$0.787N/A1.7s43 tps
DeepInfra
DeepInfra
FP8
US
$0.147$1.05$0.05250.6s36 tps
Venice
Venice
US
$0.157$1.05$0.05250.8s133 tps
AtlasCloud
AtlasCloud
FP8
US
$0.236$1.89$0.2361.3s66 tps
Parasail
Parasail
FP8
US
$0.236$1.89N/A0.7s138 tps
SiliconFlow
SiliconFlow
FP8
SG
$0.252$1.89$0.1571.1s77 tps
Alibaba
Alibaba
SG
$0.263$2.10N/A0.7s104 tps
Novita
Novita
US
$0.263$2.10N/AN/AN/A
Privacy:ZDR, prompts not retainedNo training on promptsRetention and training unclearBest latency and throughput are highlighted.
Batch API
ProviderInputOutputCached input
Auto
Autoqwen/qwen3.5-35b-a3b
$0.14/M$1.00/M$0.05/M

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 Qwen3.5 35B A3B with similar models from the same provider or model family.

Qwen3.6 35B A3B TEE

TEE/qwen3.6-35b-a3b

Qwen3.6 35B A3B is an open-weight MoE model from Alibaba's Qwen team with 35B total parameters and 3B active parameters per token. Running inside a TEE (Trusted Execution Environment), with provider attestation support.

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).

Qwen3.5 35B A3B Thinking

qwen/qwen3.5-35b-a3b:thinking

Qwen3.5 35B A3B with extended reasoning enabled. A native vision-language MoE model with hybrid attention.

Qwen3 Next 80B A3B (Instruct)

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).

Qwen3 Coder 30B A3B Instruct

qwen/qwen3-coder-30b-a3b-instruct

Qwen3 Coder 30B with 3B active parameters, optimized for code generation and technical tasks