NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/Qwen3.6 35B A3B
Alibaba Qwen logo

Qwen3.6 35B A3B

by Alibaba Qwen · Released Apr 2026

Open SourceMultimodal
42.6
avg score
Rank #185
Compare
Better than 41% of all models
Context
262K tokens (~131 books)
Input $/1M
$0.15
Output $/1M
$1.00
Type
multimodal
License
Open Source
Benchmarks
23 tested
Data as of
About

Qwen3.6-35B-A3B is an open-weight multimodal model from Alibaba Cloud with 35 billion total parameters and 3 billion active parameters per token. It uses a hybrid sparse mixture-of-experts architecture combining Gated...

Tested on 23 benchmarks · BenchGecko score 42.6. Top scores: Artificial Analysis · tau2-Bench Telecom (95.3%), OTIS Mock AIME 2024-2025 (86.7%), Artificial Analysis · GPQA Diamond (84.1%).

Looking for similar performance at lower cost?
Llama 3 8B Instruct scores 41.8 (98% as good) at $0.14/1M input · 7% cheaper
Capabilities
coding
28.8
#167 globally
math
53.5
#107 globally
knowledge
51.0
#121 globally
speed
45.0
#43 globally
general
37.5
#80 globally
Benchmark Scores
Compare All
Tested on 23 benchmarks · Ranked across 5 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
codingCompare coding →
WeirdML

Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.

34.5·
Terminal Bench

Complex terminal-based engineering tasks. Models must use command-line tools, navigate filesystems, and debug systems through shell interaction.

23.0·
mathCompare math →
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

86.7·
FrontierMath-Tiers-1-3-v2-Private
20.4·
knowledgeCompare knowledge →
GPQA diamond

Graduate-level science questions written by PhD experts. Diamond subset contains questions where experts disagree, testing deep understanding.

79.8·
Chess Puzzles

Tactical chess puzzles testing pattern recognition and multi-move calculation. Measures strategic reasoning ability.

22.1·
Quick compare:Alibaba logovs Qwen2.5 3B InstructAnthropic logovs Claude Opus 4.1OpenAI logovs o1-preview
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Model Family · Alibaba Qwen Qwen 3.6
Apr 2026
62.2
$0.32/M in262Kctx18 benchmarks
Apr 2026
42.6-19.6
$0.15/M in(-0.17)262Kctx23 benchmarks
Apr 2026
46.6+4.0
$0.19/M in(+0.04)1.0Mctx(+738K)19 benchmarks
Apr 2026
58.4+11.8
$1.03/M in(+0.84)262Kctx(-738K)13 benchmarks
Apr 2026
62.8+4.4
$0.33/M in(-0.70)1.0Mctx(+738K)25 benchmarks
Apr 2026
$0.00/M in(-0.33)1.0Mctx
Mar 2026
$0.00/M in1.0Mctx
See the full Qwen 3.6 family →
Recently Happened
Qwen3.6 35B A3B pricing increased 50%
Sep 21, 2026
Qwen3.6 35B A3B pricing dropped 29%
Aug 28, 2026
Qwen3.6 35B A3B pricing dropped 7%
Aug 17, 2026
Qwen3.6 35B A3B pricing dropped 7%
May 5, 2026
Qwen3.6 35B A3B added
Apr 27, 2026
Similar Models
Alibaba logo
Qwen2.5 3B Instruct
Alibaba
42.5n/a
Anthropic logo
Claude Opus 4.1
Anthropic
42.3$15.00/1M
OpenAI logo
o1-preview
OpenAI
43.0n/a
Links
Info
Alibaba QwenPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@Alibaba Qwen
Source Code
GitHubHugging Face
BenchGecko API
qwen3-6-35b-a3b
Specifications
  • Typemultimodal
  • Context262K tokens (~131 books)
  • ReleasedApr 2026
  • LicenseOpen Source
  • StatusActive
  • Cost / Message~$0.001
Available On
Alibaba Qwen logoAlibaba Qwen$0.15
Categories
codingmathknowledgespeedgeneral
Learn More
context-windowtransformeropen weightstokens
Share & Export
Tweet
Related Models
Alibaba logoQwen2.5 3B Instruct
42.5
Anthropic logoClaude Opus 4.1
42.3
OpenAI logoo1-preview
43.0
Alibaba Qwen logoQwen3.7 Plus
43.2
Qwen3.6 35B A3B is an open-source multimodal AI model by Alibaba Qwen, released in April 2026. It has an average benchmark score of 42.6. Context window: 262K tokens.

Related Models

Qwen2.5 3B Instruct · AlibabaClaude Opus 4.1 · Anthropico1-preview · OpenAIQwen3.7 Plus · Alibaba QwenR1 Distill Qwen 32B · DeepSeek

Benchmarks

Artificial Analysis · tau2-Bench TelecomOTIS Mock AIME 2024-2025Artificial Analysis · GPQA DiamondGPQA diamondArtificial Analysis · MMMU Pro

Related Pages

Alibaba Qwen · ProviderAlibaba Qwen · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary

Key facts · as of 2026-10-05

  • Qwen3.6 35B A3B by Alibaba Qwen. BenchGecko score 42.6, rank 185 of 312 scored models (normalized average of public benchmark scores).
  • List price $0.15 input · $1.00 output per 1M tokens (as of 2026-10-05).
  • Sold by 9 providers (as of 2026-10-05): Darkbloom (fp4) $0.0500 in / $0.70 out · AkashML (fp8) $0.10 in / $0.90 out · DeepInfra (fp8) $0.10 in / $0.95 out · Venice (fp8) $0.10 in / $1.00 out · Parasail (fp8) $0.15 in / $1.00 out · and 4 more. Every provider

How to cite · data as of 2026-10-05

Qwen3.6 35B A3B · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/qwen3-6-35b-a3b

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP