Celeris-1 vs Qwen3.8 27B (Non-reasoning)
Which is better for the money: Celeris-1 by Celeris or Qwen3.8 27B (Non-reasoning) by Alibaba? 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
Qwen3.8 27B (Non-reasoning) scores 13.9 points higher (20.2 vs 6.3) but costs 34× as much to run the benchmark ($2022 vs $60.0; flagship vs mid-range cost classes). They are not alternatives at one budget: pick Qwen3.8 27B (Non-reasoning) when you need intelligence above 6.3, and Celeris-1 when 6.3 is enough.
| Metric | Celeris-1 | Qwen3.8 27B (Non-reasoning) | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 6.3 | 20.2 | +220.6% Qwen3.8 27B (Non-reas… is 3.2× more intelligent |
| Total AA cost (lower is better) | $59.99 | $2,022.49 | +3271.4% Qwen3.8 27B (Non-reas… is 34× more expensive |
| End-to-end time (lower is better) | 0.86s | 13.98s | +1525.6% Qwen3.8 27B (Non-reas… is 1525.6% slower |
| Price per 1M tokens (in / out) | $0.20 / $0.70 | $0.50 / $3.00 | |
| Input modalities | Text, Image | Text, Image, Video |
Qwen3.8 27B (Non-reasoning) vs Celeris-1: 220.6% more intelligence, 3271.4% more cost to run the AA benchmark, and 1525.6% slower end-to-end.
Today's Pareto frontier
Against the frontier
- Celeris-1 is off the frontier: at the same or lower cost, Claude Haiku 5.5 (Medium) scores 34.50 for $56.4.
- Qwen3.8 27B (Non-reasoning) is off the frontier: at the same or lower cost, GPT-6.1 Sol (Max) scores 51.80 for $1082.
Celeris-1Qwen3.8 27B (Non-reasoning)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/celeris-1/qwen3-8-27b-non-reasoning · Markdown: /api/v1/compare/celeris-1/qwen3-8-27b-non-reasoning?format=md · llms.txt · agent skill