DeepSeek V4 Flash Vision (Max) vs Mercury 2
Which is better for the money: DeepSeek V4 Flash Vision (Max) by DeepSeek or Mercury 2 by Inception? 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
DeepSeek V4 Flash Vision (Max) scores 24.0 points higher (34.8 vs 10.8) but costs 10× as much to run the benchmark ($1081 vs $105; premium vs mid-range cost classes). They are not alternatives at one budget: pick DeepSeek V4 Flash Vision (Max) when you need intelligence above 10.8, and Mercury 2 when 10.8 is enough.
| Metric | DeepSeek V4 Flash Vision (Max) | Mercury 2 | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 34.8 | 10.8 | −69.0% Mercury 2 is 69.0% less intelligent |
| Total AA cost (lower is better) | $1,081.42 | $104.94 | −90.3% Mercury 2 is 90.3% cheaper |
| End-to-end time (lower is better) | 15.37s | 8.08s | −47.4% Mercury 2 is 47.4% faster |
| Price per 1M tokens (in / out) | $0.44 / $1.32 | $0.25 / $0.75 | |
| Input modalities | Text, Image | Text |
Mercury 2 vs DeepSeek V4 Flash Vision (Max): 69.0% less intelligence, 90.3% less cost to run the AA benchmark, and 47.4% faster end-to-end.
Today's Pareto frontier
Against the frontier
- DeepSeek V4 Flash Vision (Max) is off the frontier: at the same or lower cost, GPT-6.1 Sol (Xhigh) scores 51.00 for $662.
- Mercury 2 is off the frontier: at the same or lower cost, Claude Haiku 5.5 (High) scores 37.80 for $94.1.
DeepSeek V4 Flash Vision (Max)Mercury 2Today's Pareto 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-vision/mercury-2 · Markdown: /api/v1/compare/deepseek-v4-flash-vision/mercury-2?format=md · llms.txt · agent skill