DeepSeek V3 (Dec '24) vs Hy3
Which is better for the money: DeepSeek V3 (Dec '24) 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
Hy3 scores 16.8 points higher (25.3 vs 8.5) but costs 5.9× as much to run the benchmark ($150 vs $25.4; mid-range vs budget cost classes). They are not alternatives at one budget: pick Hy3 when you need intelligence above 8.5, and DeepSeek V3 (Dec '24) when 8.5 is enough.
| Metric | DeepSeek V3 (Dec '24) | Hy3 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 8.5 | 25.3 | +197.6% Hy3 is 3.0× more intelligent |
| Total AA cost (lower is better) | $25.37 | $149.61 | +489.7% Hy3 is 5.9× more expensive |
| End-to-end time (lower is better) | 11.89s | 33.29s | +180.0% Hy3 is 180.0% slower |
| Price per 1M tokens (in / out) | $0.32 / $0.89 | $0.14 / $0.56 | |
| Input modalities | Text | Text |
Hy3 vs DeepSeek V3 (Dec '24): 197.6% more intelligence, 489.7% more cost to run the AA benchmark, and 180.0% slower end-to-end.
Today's Pareto frontier
Against the frontier
- DeepSeek V3 (Dec '24) is off the frontier: at the same or lower cost, GPT-6 Luna (Low) scores 21.50 for $10.6.
- Hy3 is off the frontier: at the same or lower cost, GPT-6 Luna (Max) scores 38.10 for $122.
DeepSeek V3 (Dec '24)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-v3/hy3 · Markdown: /api/v1/compare/deepseek-v3/hy3?format=md · llms.txt · agent skill