Back to Explore
Qwen: Qwen3.6 35B A3B
Alibaba·Released 2026-04-27
Open Source35B262K ctxApache 2.0Multimodal
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...
Related Models
Pricing
Input
$0.14
per 1M tokens
Output
$1.00
per 1M tokens
Blended
$0.35
per 1M tokens
Cheaper than 58% of models. Median price is $0.54/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.35
Monthly
$10.65
vs. Similar Models
Qwen: Qwen3 Max ThinkingQ:+0.1
$1.56+339%
MiniMax: MiniMax M2.1Q:-0.2
$0.45+28%
Qwen3.5 397B A17B (Non-reasoning)Q:+0.4
$1.35+280%
DeepSeek V4 Pro (Non-reasoning)Q:-0.4
$0.54+53%
Performance
147
tokens/sec
Faster than 73% of models
1.36
seconds
Faster than 38% of models
38.03
seconds
Faster than 12% of models
Market Median
94 tok/s
56% faster
Median TTFT
1.11s
23% slower
Throughput/Dollar
414
tok/s per $/1M
Speed Comparison
GPT-3.5 Turbo
147 tok/s-0%
GLM-4.7-Flash (Non-reasoning)
148 tok/s+0%
Mistral Small 3
146 tok/s-1%
Context Window
262K
tokens
Larger than 62% of models
Max Output
262K
tokens
100% of context
Benchmarks
MMLU-ProNot evaluated
GPQA Diamond
84.1%
HLE
20.2%
LiveCodeBenchNot evaluated
SciCode
35.8%
TerminalBench Hard
34.8%
MATH-500Not evaluated
AIMENot evaluated
AIME 2025Not evaluated
IFBench
64.4%
Long Context Recall
63.7%
Tau2
95.3%
Market AverageTop Score
Open Source
apache-2.035BGGUF / GPTQ / AWQ
Downloads
5.6M
Likes
2.3K
VRAM (FP16)
48-80 GB
GPU
A100 80GB