DeepSeek V4 Flash 0731 (Max) vs Beam 1.0
Which is better for the money: DeepSeek V4 Flash 0731 (Max) by DeepSeek or Beam 1.0 by Reflection? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-10-24.
Clear winner
Verdict
Pick DeepSeek V4 Flash 0731 (Max)
Beam 1.0 vs DeepSeek V4 Flash 0731 (Max): 11.7% less intelligence, 59.7% more cost to run the AA benchmark, and 55.2% faster end-to-end.
| Metric | DeepSeek V4 Flash 0731 (Max) | Beam 1.0 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 34.3 | 30.3 | −11.7% Beam 1.0 is 11.7% less intelligent |
| Total AA cost (lower is better) | $199.63 | $318.73 | +59.7% Beam 1.0 is 59.7% more expensive |
| End-to-end time (lower is better) | 17.91s | 8.02s | −55.2% Beam 1.0 is 55.2% faster |
| Price per 1M tokens (in / out) | $0.14 / $0.28 | $0.40 / $1.20 | |
| Input modalities | Text | Text |
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.
- Beam 1.0 is off the frontier: at the same or lower cost, MiMo-V2.6-Pro scores 46.30 for $207.
DeepSeek V4 Flash 0731 (Max)Beam 1.0Today'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/reflection-beam · Markdown: /api/v1/compare/deepseek-v4-flash/reflection-beam?format=md · llms.txt · agent skill