Claude Sonnet 5 vs Gemini 3.1 Pro Preview for workflow automation
Claude Sonnet 5 wins, 89–83
Same 12 workflow automation tasks, marked blind by three rival labs. Claude Sonnet 5 took 7 tasks, Gemini 3.1 Pro Preview took 2, 3 tied. Tested 18 Aug 2026.
Where they differed most
A 40-person UK company wants to automate expense reimbursement end to end: receipt photographed, read, categorised, approved, paid into the employee's bank account, posted to the ledger. The finance director's brief is 'no human touches it under £100'. Go through the process and mark, for each stage, whether it can be automated, automated with a check, or must stay with a person — and give the reason in terms of what goes wrong, not in terms of policy. Name the two stages where full automation would be genuinely dangerous for a company this size, and say what the cheapest sufficient control is at each: not the most thorough control, the cheapest one that actually catches the thing.
## Stage-by-stage | Stage | Verdict | What actually goes wrong (not policy) | |---|---|---| | **1. Photograph receipt** | Automate fully | Only failure is a bad image, which shows up as a downstream OCR problem, not an approval problem. Nothing to gate here. | | **2. Read the receipt (OCR/extraction)** | Automate, with a check | OCR silently mis-reads digits (£18→£16, "1" vs "£11"), thermal paper fades, someone adds a tip in pen after the printed total. The failure is invisible unless you check it — the number just goes into the ledger wrong.…
Here is the breakdown of the end-to-end expense process for a 40-person UK company, keeping the Finance Director’s "no human touches it under £100" brief in mind. ### Part 1: The Process Stages **1. Receipt Photographed & Submitted** * **Status:** Automated. * **What goes wrong:** The employee takes a blurry photo, photographs a credit card terminal receipt instead of the itemised VAT receipt, or uploads the same receipt twice. **2. Read / OCR (Data Extraction)** * **Status:** Automated with a check. * **What goes wrong:** The soft…
Task by task
| Task | Claude Sonnet 5 | Gemini 3.1 Pro Preview |
|---|---|---|
| The MOT reminder nobody sends | 9 | 8 |
| The spec, as JSON | 9 | 8 |
| Where the human stays | 9 | 6 |
| The unhappy paths are the job | 8 | 9 |
| Do not automate this | 10 | 9 |
| It fired twice | 9 | 9 |
| The spreadsheet that runs the business | 8 | 8 |
| The alert that wakes someone up | 10 | 10 |
| How long it really takes | 9 | 8 |
| No API, no chance? | 9 | 8 |
| Rules, not guesses | 9 | 8 |
| Refuse the scraper | 8 | 9 |
Full receipts: Claude Sonnet 5, Gemini 3.1 Pro Preview · judges gemini-3.1-pro-preview, gpt-5.6-terra, grok-4.5
Questions people ask
Which is better for workflow automation: Claude Sonnet 5 or Gemini 3.1 Pro Preview?
Claude Sonnet 5 — it scored 89/100 against 83/100 on our 12-task workflow automation 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 workflow automation 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 workflow automation head-to-heads: Claude Sonnet 5 vs GPT-5.6 Sol · Gemini 3.1 Pro Preview vs GPT-5.6 Sol · Claude Sonnet 5 vs GPT-5.5 · Claude Sonnet 5 vs Grok 4.5 · Claude Sonnet 5 vs Kimi K3 · Claude Opus 4.8 vs Claude Sonnet 5
Full ranking: Best AI for workflow automation · model pages: Claude Sonnet 5, Gemini 3.1 Pro Preview