# Asked the same question twice, the models moved their own first choice 70% of the time. The leader still held in 7 of 10 categories.

HR AI Recommendation Index, September 2026 Edition. Page: https://hr-ai-index.com/research/repeat-measurement/

Ten categories were asked again one days after the edition run, with nothing changed: the same prompts, the same model versions, the same settings, one buyer segment, six framings, twelve models, 720 answers. Whatever differs between the two runs is noise, and the noise is what this page measures. That is the reason the index asks every question six ways of every model rather than once: on its own, one ask moved 70% of the time; seventy-two, read together, are what the index reports.

## Model by model

- MiniMax M2.5: 89% of 38 pairs flipped
- Qwen 3.7 Flash: 83% of 48 pairs flipped
- GPT-5.4 mini: 76% of 38 pairs flipped
- GLM 4.7 FlashX: 76% of 38 pairs flipped
- Kimi K2: 72% of 40 pairs flipped
- DeepSeek V4 Flash: 70% of 47 pairs flipped
- Gemini 3.5 Flash: 69% of 48 pairs flipped
- Mistral Small: 68% of 37 pairs flipped
- Claude Haiku 4.5: 67% of 36 pairs flipped
- Llama 4 Maverick: 55% of 31 pairs flipped
- Grok 4.1 Fast: 55% of 44 pairs flipped
- Perplexity Sonar: 51% of 35 pairs flipped

## Category by category

| Category | Leader | Share, run one | Share, repeat | Move | Leader |
|---|---|---|---|---|---|
| Contingent workforce man | SimpleVMS | 35% | 22% | -13 points | held |
| Compensation benchmarkin | Payscale | 24% | 33% | 9 points | held |
| AI recruiting assistants | GoPerfect | 19% | 11% | -8 points | changed: Workable |
| Employee onboarding | BambooHR | 29% | 25% | -4 points | changed: Rippling |
| Benefits administration | PlanSource | 23% | 20% | -2 points | changed: Gusto |
| Learning management syst | TalentLMS | 48% | 50% | 2 points | held |
| Performance management | Lattice | 41% | 43% | 2 points | held |
| Employee engagement surv | Culture Amp | 40% | 42% | 2 points | held |
| Global payroll | Deel | 22% | 24% | 2 points | held |
| Coaching platforms | Growthspace | 26% | 27% | 1 point | held |

Share floor: 9 points (the 90th percentile); median move 2 points; largest 13 points. Pooled flip rate 70%.

Published under CC BY 4.0. The output is the models' output; nothing here is a recommendation by the index.