Claude Opus 4.8 vs DeepSeek V4 Pro for multi-step research

The verdict

Claude Opus 4.8 wins, 8479

Same 12 multi-step research tasks, marked blind by three rival labs. Claude Opus 4.8 took 7 tasks, DeepSeek V4 Pro took 2, 3 tied. Tested 18 Aug 2026.

Where they differed most

The task — The disagreement is definitional

Three sources on the average deposit a UK first-time buyer needs. Source A — a high-street lender, 2025: '£53,414', based on mortgages it approved that year. Source B — a building society, 2025: '£34,500', first-time buyers outside London. Source C — a London estate agency, 2024: '£71,000'. These are usually written up as a contradiction. Show that they are not: name the hidden variable inside each figure (who is counted, where, when, and what counts as a deposit), and explain which comparison would be needed before any two of them could disagree. Then write the one sentence about first-time-buyer deposits that is true given all three, and name the thing a specific reader — a couple saving in Derby — still cannot learn from any of them.

## The hidden variable in each figure

**Source A — £53,414 (high-street lender, 2025).**
Who is counted: only people this *specific lender* actually approved for a mortgage — i.e. buyers who already had enough deposit to clear that lender's own criteria. This is a *selected*, successful population, not all first-time buyers. Where: wherever this lender happens to lend, weighted toward wherever it approves most volume. What counts as a deposit: the cash the buyer put down on the deals that completed. The hidden variable is **survivorship** — it…
They are not contradictions; they are averages of different groups, places, times and definitions.

| Source | Who is counted | Where | When | What counts as a deposit |
|---|---|---|---|---|
| **A: High-street lender, £53,414** | First-time buyers approved for mortgages by that specific lender | UK-wide/lender’s national mortgage book, including London | 2025 | Mortgage deposit at completion: purchase price minus mortgage; may include gifted funds; not a minimum required |
| **B: Building society, £34,500** | First-time buyer mortgage customer…

Task by task

TaskClaude Opus 4.8DeepSeek V4 Pro
Four numbers, one high street98
The trial that both worked and didn't98
Break the brief into answerable steps109
Corroboration that isn't88
The disagreement is definitional83
A forecast is not an outcome108
The evidence table, as JSON68
Funded, and still the best data98
Write only the limits section88
Not answerable from what you have1010
The statistic and the doorstep98
Refuse the manufactured briefing59

Full receipts: Claude Opus 4.8, DeepSeek V4 Pro · judges gemini-3.1-pro-preview, gpt-5.6-terra, grok-4.5

Questions people ask

Which is better for multi-step research: Claude Opus 4.8 or DeepSeek V4 Pro?

Claude Opus 4.8 — it scored 84/100 against 79/100 on our 12-task multi-step research suite, winning 7 tasks to 2 with 3 tied. Every answer was marked blind by three judges from three rival AI labs.

How was this tested?

Both models answered the identical published multi-step research tasks. Three AI judges from three different labs scored every answer blind against a fixed rubric; mechanically checkable rules (word limits, banned phrases) are enforced by the test harness in code. The raw outputs and judge verdicts are downloadable.

More multi-step research head-to-heads: Claude Opus 4.8 vs Claude Sonnet 5 · Claude Sonnet 5 vs DeepSeek V4 Pro · Claude Opus 4.8 vs GPT-5.5 · DeepSeek V4 Pro vs GPT-5.5 · Claude Opus 4.8 vs GPT-5.6 Sol · DeepSeek V4 Pro vs GPT-5.6 Sol

Full ranking: Best AI for multi-step research · model pages: Claude Opus 4.8, DeepSeek V4 Pro