Qwen3.7 Max (NovitaAI)

Alibaba’s Qwen3.7 Max is positioned as a proprietary reasoning model for agentic work rather than ordinary chat. It is designed for extended, multi-step tool use, including software engineering, terminal tasks, document handling, and other workflows that require sustained context. The model’s large one-the cataloged API limit context supports information-heavy projects, while its instruction-following and reasoning results indicate strong performance on structured tasks.

Benchmark evidence shows particular strength in coding and tool-oriented evaluations, including LiveCodeBench, Terminal-Bench, SWE-bench, MCPMark, and Claw-Eval, alongside strong results in mathematics, multilingual knowledge, long-context retrieval, and professional reasoning. These results make it a practical fit for coding agents and research or document workflows that need iterative tool use and long context. Its performance is less consistent on some open-ended professional, scientific, game-playing, and creative tasks, so workload-specific evaluation is advisable.

LLM Gatewaynovita/qwen3.7-maxqwen

Open provider page Provider docs

Quick Info

Provider: LLM Gateway
Model key: novita/qwen3.7-max
Release date: May 21, 2026
Last updated: May 21, 2026
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $1.25
Output token cost: $3.75

Limits

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

Transparent token rates

Compare Qwen3.7 Max (NovitaAI) 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 (NovitaAI)

LLM Gateway

CoverageBenchmark

Third-party review explicitly naming Qwen3.7-Max with concrete launch and capability detail. Reports a May 20, 2026 launch at the Alibaba Cloud Summit in Hangzhou, an Artificial Analysis Intelligence Index score of 56.6 at launch (top 10 of 151 measured models), a 1 million-token context window with max output of 65,53 The review does not mention NovitaAI or llmgateway.io, so the article serves strictly as model-level coverage of the underlying Alibaba Qwen3.7 Max model rather than any gateway-specific claims. Some pricing figures in the article ($2.50 input / $7.50 output per 1M tokens) conflict with other aggregator sources ($1.25/

LLM Gateway

CoverageBenchmark

Aggregator benchmark listing for Qwen3.7 Max showing 95 evaluation rows attributed to Qwen, with the most concrete benchmark breakdown among the candidates. Leaderboard entries include MCPMark at 60.8% (rank 1), Terminal-Bench Hard at 50.8% (rank 10, dated 2026-06-10), VitaBench Cross-Scenarios Avg@4 at 47.9% (rank 2), The page does not mention NovitaAI or any gateway/reseller hosting context, so it is treated strictly as model-level material about the Qwen3.7 Max family member created by Qwen. Entries span late May through mid-June 2026, giving a temporally grounded snapshot of where the model stood on agentic and tool-use benchmark