Gemini 3.1 Pro Preview vs Gemini 3.5 Flash for multi-step research
Gemini 3.1 Pro Preview wins, 82–78
Same 12 multi-step research tasks, marked blind by three rival labs. Gemini 3.1 Pro Preview took 3 tasks, Gemini 3.5 Flash took 1, 8 tied. Tested 18 Aug 2026.
Where they differed most
A charity trustee asks: 'should our shop open on Sundays?' That is a decision, not a research question. Turn it into exactly 5 sub-questions, numbered 1-5, one line each and no sub-points. Each line must name (a) the sub-question, (b) the kind of source or record that would answer it — the shop's own till data, the rota and wage cost, the lease or planning conditions, comparable local shops, the volunteer pool — and (c) the finding that would flip the decision. No general research advice: five specific, answerable questions.
1. What is the projected Sunday revenue based on the shop's own till data, where the decision flips to 'no' if projected takings fall below the daily break-even threshold. 2. How much will Sunday operations cost based on the rota and wage cost, where the decision flips to 'no' if staffing expenses exceed the projected Sunday gross profit. 3. Are we legally permitted to trade on Sundays according to the lease or planning conditions, where the decision flips to 'no' if there is an explicit prohibition on Sunday opening. 4. Is there an active Sund…
1. Are we legally permitted to trade on Sundays? Check the lease or planning conditions; a finding of an explicit legal prohibition or restrictive covenant would immediately veto the opening. 2. Will our existing volunteers cover Sunday shifts? Survey the volunteer pool; a finding that fewer than the minimum required volunteers are willing to work Sundays would prevent us from opening. 3. What will it cost in paid staff wages to run the Sunday shift? Analyze the rota and wage cost; a finding that Sunday premium pay rates exceed the projected…
Task by task
| Task | Gemini 3.1 Pro Preview | Gemini 3.5 Flash |
|---|---|---|
| Four numbers, one high street | 8 | 8 |
| The trial that both worked and didn't | 8 | 8 |
| Break the brief into answerable steps | 10 | 8 |
| Corroboration that isn't | 8 | 8 |
| The disagreement is definitional | 8 | 6 |
| A forecast is not an outcome | 8 | 8 |
| The evidence table, as JSON | 6 | 7 |
| Funded, and still the best data | 7 | 7 |
| Write only the limits section | 9 | 9 |
| Not answerable from what you have | 10 | 9 |
| The statistic and the doorstep | 8 | 8 |
| Refuse the manufactured briefing | 8 | 8 |
Full receipts: Gemini 3.1 Pro Preview, Gemini 3.5 Flash · judges claude-sonnet-5, gpt-5.6-terra, grok-4.5
Questions people ask
Which is better for multi-step research: Gemini 3.1 Pro Preview or Gemini 3.5 Flash?
Gemini 3.1 Pro Preview — it scored 82/100 against 78/100 on our 12-task multi-step research suite, winning 3 tasks to 1 with 8 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 Sonnet 5 vs Gemini 3.1 Pro Preview · Claude Sonnet 5 vs Gemini 3.5 Flash · Gemini 3.1 Pro Preview vs GPT-5.5 · Gemini 3.5 Flash vs GPT-5.5 · Gemini 3.1 Pro Preview vs GPT-5.6 Sol · Gemini 3.5 Flash vs GPT-5.6 Sol
Full ranking: Best AI for multi-step research · model pages: Gemini 3.1 Pro Preview, Gemini 3.5 Flash