Qwen 3.5 27B
QwenBalanced
ThinkingTool UseVisionStructured Output
About this model
The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance. Its overall capabilities are comparable to those of the Qwen3.5-122B-A10B.
Performance Tier
Balanced
Qwen 3.5 27B is a balanced model from Qwen : strong performance at a reasonable price.
Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.
Pricing
This model is included in Elosia plans| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.195 |
| Output (completion) | $1.56 |
Capabilities
Context Length262K
Max Output Tokens66K
TokenizerQwen3
Inputtext, image, video
Outputtext
Release DateFebruary 25, 2026
Benchmarks
General Intelligence
MMLU
Not reported
MMLU-Pro
86.1%
GPQA Diamond
85.5%
Mathematics
MATH-500
Not reported
Programming
HumanEval
Not reported
SWE-bench Verified
72.4%
Reasoning
IFEval
95%
Humanity's Last Exam
48.5%
Multimodal
MMMU-Pro
75%
Recommended Use Cases
CodingAnalysisResearchGeneral ChatTranslation
Strengths
- Strongest dense model in the Qwen 3.5 medium range
- Excellent instruction following (IFEval 95.0%)
- Top-tier software engineering (SWE-bench 72.4%)
- Native multimodal support (images, video, documents)
- 262K native context, extensible to 1M with YaRN
Limitations
- Higher cost than Flash variant ($0.27/M vs $0.10/M input)
- Dense architecture requires more compute than MoE alternatives
- No published MMLU, HumanEval, or MATH-500 scores
Resources
This model may use your data for training
