GPT-6.1 Terra (Non-reasoning) vs Hy3
Which is better for the money: GPT-6.1 Terra (Non-reasoning) by OpenAI or Hy3 by Tencent? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-27.
Trade-off
Verdict
Depends on your budget
GPT-6.1 Terra (Non-reasoning) scores 8.2 points higher (33.5 vs 25.3) but costs 1.0× as much to run the benchmark ($154 vs $150; both mid-range class). Pick GPT-6.1 Terra (Non-reasoning) when the extra intelligence is worth the premium, otherwise Hy3.
| Metric | GPT-6.1 Terra (Non-reasoning) | Hy3 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 33.5 | 25.3 | −24.5% Hy3 is 24.5% less intelligent |
| Total AA cost (lower is better) | $153.58 | $150.27 | −2.2% Hy3 is 2.2% cheaper |
| End-to-end time (lower is better) | 6.51s | 33.53s | +415.1% Hy3 is 415.1% slower |
| Price per 1M tokens (in / out) | $1.00 / $5.00 | $0.14 / $0.56 | |
| Input modalities | Text, Image | Text |
Hy3 vs GPT-6.1 Terra (Non-reasoning): 24.5% less intelligence, 2.2% less cost to run the AA benchmark, and 415.1% slower end-to-end.
Today's Pareto frontier
Against the frontier
- GPT-6.1 Terra (Non-reasoning) is off the frontier: at the same or lower cost, GPT-6 Luna (Max) scores 38.10 for $122.
- Hy3 is off the frontier: at the same or lower cost, GPT-6 Luna (Max) scores 38.10 for $122.
GPT-6.1 Terra (Non-reasoning)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/gpt-6-1-terra-non-reasoning/hy3 · Markdown: /api/v1/compare/gpt-6-1-terra-non-reasoning/hy3?format=md · llms.txt · agent skill