Qwen3.7 Max

Qwen3.7-Max is positioned by its creators as a proprietary foundation for the agent era, designed to keep coherent behavior across very long task horizons. The release narrative emphasizes coding work ranging from frontend prototyping to multi-file engineering, along with office and productivity assistance through MCP integrations and multi-agent orchestration. As a flagship in the Qwen3.7 series, it is intended to generalize across agent scaffolds so that the same underlying model can be deployed through different frameworks without losing reliability.

In benchmark listings, the model records a coding index near 66 percent and a strong graduate-level reasoning score, while a composite intelligence index places it in the mid-forties, suggesting balanced rather than purely specialized capability. It supports reasoning, tool calling, prompt caching, and structured JSON output, making it well suited to autonomous loops that revisit the same context repeatedly. For teams building coding assistants, workflow automators, or sustained agent pipelines, Qwen3.7-Max offers a versatile core that pairs long-context retention with practical productivity features.

ClinePasscline-pass/qwen3.7-maxqwen

Open provider page Provider docs

Quick Info

Provider: ClinePass
Model key: cline-pass/qwen3.7-max
Release date: May 21, 2026
Last updated: May 21, 2026
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $2.50
Output token cost: $7.50

Limits

Output tokens: 65,536 tokens
Context window: 1,000,000 tokens

Transparent token rates

Compare Qwen3.7 Max pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Qwen3.7 Max

ClinePass

Coverage

Explainx.ai's blog post dated May 22, 2026 (with a July 19, 2026 update banner) confirms Alibaba Cloud unveiled Qwen 3.7-Max on May 20, 2026 as a "versatile agent foundation" designed for sustained autonomous execution rather than chasing general-purpose benchmark ceilings. The article frames the release as a shift tow The piece includes a coding-agent benchmark table comparing Qwen 3.7-Max against Opus 4.6 Max and DeepSeek V4 Pro: Terminal Bench 2.0 69.7 (vs 65.4 / 67.9), SWE-bench Verified 80.4 (vs 80.8 / 80.6), SWE-bench Pro 60.6 (vs 57.3 / 59.0), and SciCode 53.5 (vs 51.9 / not reported). It also explains Terminal Bench 2.0-Termi

ClinePass

CoverageBenchmark

DataCamp's May 22, 2026 explainer reports that Alibaba's Qwen team released Qwen3.7-Max as a proprietary flagship positioned as a "versatile agent foundation" above Qwen3.6-Plus, available through Alibaba Cloud Model Studio with a 1M-token context window and text-only I/O. The article cites concrete benchmark numbers: The same article flags notable verbosity: Artificial Analysis observed approximately 97 million tokens generated during evaluation, far above the 24M median, with cost implications for long agentic sessions. Improvements over Qwen3.6-Plus are described as "substantial across every category," with the widest gaps on lon

ClinePass

Coverage

ApXML's model card lists Qwen3.7 Max as closed-source, proprietary, dense architecture, text-only modality, with a 1M-token context length and a release date of May 19, 2026. The card describes it as "Alibaba's proprietary frontier model for the agent era," purpose-built as a versatile agent foundation capable of writi Leaderboard metadata on the same page places Qwen3.7 Max at overall rank 104 and coding rank 72, with category scores including Text Arena 1474 (#26), WebDev Arena 1517 (#28), LiveBench Global 0.73 (#27), LiveBench Reasoning 0.83 (#27), LiveBench Data Analysis 0.72 (#28), LiveBench Coding 0.74 (#32), LiveBench Agentic