Back to Explore
Qwen: Qwen3.6 27B
Alibaba·Released 2026-04-27
Open Source27B262K ctxApache 2.0Multimodal
About
Qwen3.6 27B is a dense 27-billion-parameter language model from the Qwen Team at Alibaba, released in April 2026. It features hybrid multimodal capabilities — accepting text, image, and video inputs...
Related Models
Pricing
Input
$0.26
per 1M tokens
Output
$2.39
per 1M tokens
Blended
$0.79
per 1M tokens
Cheaper than 43% of models. Median price is $0.54/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.79
Monthly
$23.73
vs. Similar Models
xAI: Grok 4.20Q:-0.1
$1.56+98%
DeepSeek V4 Flash (Reasoning, High Effort)Q:+0.3
$0.17-78%
xAI: Grok 4.3Q:+0.5
$1.56+98%
Grok 4.20 0309 (Reasoning)Q:-0.6
$3.00+279%
Performance
60
tokens/sec
Faster than 27% of models
1.45
seconds
Faster than 35% of models
96.24
seconds
Faster than 0% of models
Market Median
94 tok/s
37% slower
Median TTFT
1.11s
31% slower
Throughput/Dollar
76
tok/s per $/1M
Speed Comparison
Gemma 4 E4B (Non-reasoning)
60 tok/s-0%
Qwen3 8B (Reasoning)
60 tok/s+0%
Gemma 4 E4B (Reasoning)
60 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
84.2%
HLE
21.6%
LiveCodeBenchNot evaluated
SciCode
39.8%
TerminalBench Hard
34.8%
MATH-500Not evaluated
AIMENot evaluated
AIME 2025Not evaluated
IFBench
67.6%
Long Context Recall
68.7%
Tau2
94.2%
Market AverageTop Score
Open Source
apache-2.027BGGUF / GPTQ / AWQ
Downloads
5.5M
Likes
1.8K
VRAM (FP16)
24-48 GB
GPU
A6000 / M3 Ultra