DeepSeek V4 Flash 0731 (Max) vs Hy3
Which is better for the money: DeepSeek V4 Flash 0731 (Max) by DeepSeek or Hy3 by Tencent? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-24.
Trade-off
Verdict
Depends on your budget
DeepSeek V4 Flash 0731 (Max) scores 9.0 points higher (34.3 vs 25.3) but costs 1.3× as much to run the benchmark ($200 vs $150; both mid-range class). Pick DeepSeek V4 Flash 0731 (Max) when the extra intelligence is worth the premium, otherwise Hy3.
| Metric | DeepSeek V4 Flash 0731 (Max) | Hy3 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 34.3 | 25.3 | −26.2% Hy3 is 26.2% less intelligent |
| Total AA cost (lower is better) | $199.63 | $149.61 | −25.1% Hy3 is 25.1% cheaper |
| End-to-end time (lower is better) | 17.91s | 33.29s | +85.9% Hy3 is 85.9% slower |
| Price per 1M tokens (in / out) | $0.14 / $0.28 | $0.14 / $0.56 | |
| Input modalities | Text | Text |
Hy3 vs DeepSeek V4 Flash 0731 (Max): 26.2% less intelligence, 25.1% less cost to run the AA benchmark, and 85.9% slower end-to-end.
Today's Pareto frontier
Against the frontier
- DeepSeek V4 Flash 0731 (Max) is off the frontier: at the same or lower cost, Claude Haiku 5.5 (Xhigh) scores 41.20 for $158.
- Hy3 is off the frontier: at the same or lower cost, GPT-6 Luna (Max) scores 38.10 for $122.
DeepSeek V4 Flash 0731 (Max)Hy3Today's Pareto frontierOpen-weights frontierEvery other model
Matchup
Compare something else
Agents:
Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/compare/deepseek-v4-flash/hy3 · Markdown: /api/v1/compare/deepseek-v4-flash/hy3?format=md · llms.txt · agent skill