Hy3vsQwen3.7 Max
Across 7 shared benchmarks, Qwen3.7 Max leads overall: Hy3 wins 1, Qwen3.7 Max wins 5, with 1 ties and an average score difference of -1.03.
Hy3
腾讯AI实验室 · 2026-07-06 · Reasoning model
Qwen3.7 Max
阿里巴巴 · 2026-05-20 · Reasoning model
Hy31 win(14%)Ties1(71%)5 winsQwen3.7 Max
Benchmark scores
Grouped by capability, sorted by largest gap within each. 7 shared benchmarks.
Coding and Software Engineer
Qwen3.7 Max 3/3| Benchmark | Hy3 | Qwen3.7 Max | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 57.9018 / 57Thinking High (With Tools) | 60.6012 / 57Thinking (With Tools) | -2.70 |
| SWE-bench Multilingual | 75.808 / 25Thinking High (With Tools) | 78.304 / 25Thinking (With Tools) | -2.50 |
| SWE-bench Verified | 7824 / 114Thinking High (With Tools) | 80.4013 / 114Thinking (With Tools) | -2.40 |
AI Agent - Tool Usage
Hy3 1/1| Benchmark | Hy3 | Qwen3.7 Max | Diff |
|---|---|---|---|
| MCP-Atlas | 79.109 / 38Thinking High (With Tools) | 76.4016 / 38Thinking (With Tools) | +2.70 |
General Evaluation
Qwen3.7 Max 1/1| Benchmark | Hy3 | Qwen3.7 Max | Diff |
|---|---|---|---|
| GPQA Diamond | 90.4036 / 226Thinking High (No Tools) | 92.4022 / 226最高（无工具） | -2 |
General Knowledge
Qwen3.7 Max 1/1| Benchmark | Hy3 | Qwen3.7 Max | Diff |
|---|---|---|---|
| HLE | 53.2019 / 181Thinking High (With Tools) | 53.5018 / 181Thinking (With Tools) | -0.30 |
Math and Reasoning
Even 1/1| Benchmark | Hy3 | Qwen3.7 Max | Diff |
|---|---|---|---|
| IMO-AnswerBench | 903 / 23Thinking High (No Tools) | 903 / 23最高（无工具） | — |
Specs
| Field | Hy3 | Qwen3.7 Max |
|---|---|---|
| Publisher | 腾讯AI实验室 | 阿里巴巴 |
| Release date | 2026-07-06 | 2026-05-20 |
| Model type | Reasoning model | Reasoning model |
| Architecture | MoE | Dense |
| Parameters | 295B | Not available |
| Context length | 256K | 1M |
| Max output | Not available | 64K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Hy3 | Qwen3.7 Max |
|---|---|---|
| Text input | ¥1.2 / 1M tokens | ¥12 / 1M tokens |
| Text output | ¥4 / 1M tokens | ¥36 / 1M tokens |
| Cache read | ¥0.4 / 1M tokens | Not public |
Summary
- Hy3leads in:AI Agent - Tool Usage (1/1)
- Qwen3.7 Maxleads in:Coding and Software Engineer (3/3), General Evaluation (1/1), General Knowledge (1/1)
- Tied in:Math and Reasoning
On average across the 7 shared benchmarks, Qwen3.7 Max scores 1.03 higher.
Largest single-benchmark gap: MCP-Atlas — Hy3 79.10 vs Qwen3.7 Max 76.40 (+2.70).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.