Qwen 3.7 Plus
QwenBalanced
ThinkingTool UseVisionStructured Output
About this model
Qwen3.7-Plus is a cost-effective model in Alibaba's Qwen3.7 series. It supports text and image input with text output, building on the series' text capabilities with a comprehensive upgrade to its...
Performance Tier
Balanced
Qwen 3.7 Plus 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 plansAffordable
Low cost. Suitable for sustained use and high-volume interactions.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.320 |
| Output (completion) | $1.28 |
| Cache read | $0.064 |
| Cache write | $0.400 |
Capabilities
Context Length1.0M
Max Output Tokens66K
TokenizerQwen
Inputtext, image
Outputtext
Release DateJune 3, 2026
Benchmarks
General Intelligence
MMLU
Not reported
GPQA Diamond
90.3%
Mathematics
MATH-500
Not reported
AIME 2025
93.3%
Programming
HumanEval
Not reported
Agentic
SWE-bench Pro
57.6%
Recommended Use Cases
CodingAnalysisResearchData Extraction
Strengths
- Multimodal agent variant of the Qwen3.7 line — native text, image and video input for scene perception, screen reading and GUI control
- 1M-token context window for full-codebase and long-document agentic workflows, with strong long-context retrieval
- Math reasoning on par with the Qwen3.7 Max flagship (AIME 2025 93.3, independently verified) at a fraction of the cost
- Low cost for its class at $0.32/M input and $1.28/M output — roughly 60% cheaper input than the Max tier
- Strong GUI / computer-use agent profile — screen understanding, tool use and code-from-visual-reference
Limitations
- Proprietary and closed-weight — no downloadable checkpoint, unlike Qwen's open-weight tradition; API-only deployment
- Sits below the Qwen3.7 Max tier on hard text benchmarks (GPQA Diamond 90.3 vs Max 92.4)
- Standard text suites largely unpublished and benchmarks mostly self-reported — independent verification limited; high output verbosity noted
Resources
This model may use your data for training
