Hy3 vs Muse Spark 1.2 (Xhigh)
Which is better for the money: Hy3 by Tencent or Muse Spark 1.2 (Xhigh) by Meta? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-25.
Trade-off
Verdict
Depends on your budget
Muse Spark 1.2 (Xhigh) scores 14.3 points higher (39.6 vs 25.3) but costs 9.3× as much to run the benchmark ($1385 vs $150; premium vs mid-range cost classes). They are not alternatives at one budget: pick Muse Spark 1.2 (Xhigh) when you need intelligence above 25.3, and Hy3 when 25.3 is enough.
| Metric | Hy3 | Muse Spark 1.2 (Xhigh) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 25.3 | 39.6 | +56.5% Muse Spark 1.2 (Xhigh) is 56.5% more intelligent |
| Total AA cost (lower is better) | $149.61 | $1,385.40 | +826.0% Muse Spark 1.2 (Xhigh) is 9.3× more expensive |
| End-to-end time (lower is better) | 33.13s | 20.85s | −37.1% Muse Spark 1.2 (Xhigh) is 37.1% faster |
| Price per 1M tokens (in / out) | $0.14 / $0.56 | $1.25 / $4.25 | |
| Input modalities | Text | Text, Image, Speech, Video |
Muse Spark 1.2 (Xhigh) vs Hy3: 56.5% more intelligence, 826.0% more cost to run the AA benchmark, and 37.1% faster end-to-end.
Today's Pareto frontier
Against the frontier
- Hy3 is off the frontier: at the same or lower cost, GPT-6 Luna (Max) scores 38.10 for $122.
- Muse Spark 1.2 (Xhigh) is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
Hy3Muse Spark 1.2 (Xhigh)Today'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/hy3/muse-spark-1-2 · Markdown: /api/v1/compare/hy3/muse-spark-1-2?format=md · llms.txt · agent skill