ModelBeat
Models/Qwen 3.6 35B-A3B
All models
A

Alibaba: Qwen 3.6 35B-A3B

qwen/qwen3.6-35b-a3b

Qwen 3.6 35B-A3B is Alibaba’s open-weights AI model, released Apr 14, 2026. It costs $0.10/M input and $0.90/M output tokens, and handles a 262K-token context window.

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.

Context
262K
Input
$0.10 / 1M
Output
$0.90 / 1M
Released
Apr 14, 2026
Modalities
textimagevideotext

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Qwen 3.6 35B-A3B lands among all models tracked on Model Beat.

29.0
Intelligence Index
Epoch AI
29th percentile of tracked models
10.0
Coding Index
Epoch AI
10th percentile of tracked models
45.0
Agentic Index
Epoch AI
45th percentile of tracked models

Reasoning

3 evals
GPQA Diamond84.8%

Graduate-level scientific reasoning

Humanity's Last Exam22.2%

Frontier of human expert knowledge

WeirdML34.5%

Novel ML problem-solving

Coding

2 evals
SciCode36.6%

Scientific research coding

WebDev Arena1250

Human-rated web development

Math

1 evals
AIME 2024/202586.7%

Olympiad-qualifier math

Agentic & Tools

2 evals
Terminal-Bench23.0%

Command-line agentic tasks

τ²-bench95.3%

Tool-agent-user reliability

In the news

4

Changelog

7
  • Qwen 3.6 35B-A3B: cheapest credible provider via AkashML now $0.9 per 1M output tokens (was $0.95, -5%)cheapest provider · Aug 30, 2026
  • Qwen 3.6 35B-A3B: reported Terminal-Bench score fell from 24.6% to 23.0% (-7%)benchmark · Aug 26, 2026
  • Qwen 3.6 35B-A3B: Humanity's Last Exam improved from 20.2% to 22.2% (+10%)benchmark · Aug 6, 2026
  • Qwen 3.6 35B-A3B: cheapest credible provider via DeepInfra now $0.95 per 1M output tokens (was $1, -5%)cheapest provider · Jul 31, 2026
  • Qwen 3.6 35B-A3B: cheapest credible provider via Venice now $0.1 per 1M input tokens (was $0.14, -29%)cheapest provider · Jul 29, 2026
  • Qwen 3.6 35B-A3B: cheapest credible provider via AkashML now $0.14 per 1M input tokens (was $0.13, +8%)cheapest provider · Jul 26, 2026
  • Qwen 3.6 35B-A3B: cheapest credible provider via DekaLLM now $0.13 per 1M input tokens (was $0.14, -7%)cheapest provider · Jul 16, 2026

Compare Qwen 3.6 35B-A3B with

Frequently asked questions

What is Qwen 3.6 35B-A3B?

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.

Who created Qwen 3.6 35B-A3B?

Qwen 3.6 35B-A3B is an AI model developed by Alibaba, part of the Qwen family, released Apr 14, 2026.

How much does Qwen 3.6 35B-A3B cost?

Qwen 3.6 35B-A3B is served from $0.10 per million input tokens and $0.90 per million output tokens (cheapest credible provider on OpenRouter: AkashML).

What is Qwen 3.6 35B-A3B's context window?

Qwen 3.6 35B-A3B supports a context window of up to 262K tokens.

How does Qwen 3.6 35B-A3B perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Qwen 3.6 35B-A3B scores 95.3% on τ²-bench, 84.8% on GPQA Diamond, 22.2% on Humanity's Last Exam.

Is Qwen 3.6 35B-A3B open source?

Qwen 3.6 35B-A3B is released with open weights (Open weights (unrestricted)), so the model can be downloaded and self-hosted.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Sep 17, 2026.