# Compensation benchmarking data: what AI models recommend, November 2026

Gane Index, November 2026 Edition. Asked as "salary benchmarking data provider" and as "compensation survey data", six framings each, to fourteen models with search on, on behalf of a mid-market B2B company. Page: https://www.gane.ai/index/hr/rewards/compensation-benchmarking-data

**Standing:** Payscale leads with 18% of first choices; verdict contested. 55 first choices across the direct, paraphrase, budget and scale prompts.

## First-choice share

| Rank | Product | Share | Negative rate | Labels |
|---|---|---|---|---|
| 1 | [Payscale](https://www.gane.ai/index/hr/vendors/payscale/index.md) | 18% | 9% | 32 |
| 2 | [Pave](https://www.gane.ai/index/hr/vendors/pave/index.md) | 15% | 0% | 36 |
| 3 | [Comprehensive.io](https://www.gane.ai/index/hr/vendors/comprehensive-io/index.md) | 9% | 0% | 15 |
| 5 | [Salary.com CompAnalyst](https://www.gane.ai/index/hr/vendors/salary-com-companalyst/index.md) | 4% | 14% | 22 |
| 8 | [Glassdoor](https://www.gane.ai/index/hr/vendors/glassdoor/index.md) | 4% | 38% | 16 |
| 9 | [Mercer Total Remuneration Survey](https://www.gane.ai/index/hr/vendors/mercer-total-remuneration-survey/index.md) | 4% | 50% | 34 |
| 27 | [Radford](https://www.gane.ai/index/hr/vendors/radford/index.md) | 2% | 48% | 21 |
| 28 | [Willis Towers Watson](https://www.gane.ai/index/hr/vendors/willis-towers-watson/index.md) | 2% | 46% | 13 |
| 29 | [Ravio](https://www.gane.ai/index/hr/vendors/ravio/index.md) | 0% | 0% | 18 |

Rank is by share of first choices among all 135 products named; products with the same share are tied, and the order among them carries no meaning. The table lists those with at least 10 labels.

## Each model's first choice on the direct prompt

- Claude Haiku 4.5: Pave, SalaryCube; alternatives CompUp, OpenComp, Payscale Ascent
- GPT-5.4 mini: Salary.com CompAnalyst; alternatives Compensation-IQ, Pave, Payscale Ascent
- Gemini 3.5 Flash: Pave; alternatives Carta Total Compensation, Payscale, Salary.com CompAnalyst
- Perplexity Sonar: Payscale; alternatives Pave, Ravio, SalaryCube
- Grok 4.20: Payscale; alternatives Pave, Ravio, Salary.com CompAnalyst
- Mistral Small: Pave; alternatives Mercer Total Remuneration Survey, Payscale
- DeepSeek V4 Flash: Payscale; alternatives Mercer Benchmark Database, Pave, Ravio, SalaryCube
- Llama 4 Maverick: SalaryCube; alternatives Mercer Total Remuneration Survey, Pave, Payscale
- Qwen 3.7 Flash: Comprehensive.io; alternatives Figures, Pave, Payscale, Salary.com CompAnalyst
- Kimi K2: Payscale; alternatives Mercer Total Remuneration Survey, Pave, Radford, Salary.com CompAnalyst
- GLM 4.7 FlashX: Figures, SalaryCube Comp Planning; alternatives Pave, Salary.com CompAnalyst
- MiniMax M2.5: CompAnalyst, Payscale; alternatives Pave
- GPT-6 Luna: Payscale; alternatives Pave
- Muse Glimmer 30B: Payscale; alternatives CompAnalyst, Comprehensive.io, Pave, Salary.com CompAnalyst

## Sources the answers cite

67 of 84 answers came back with a source list, from 14 of 14 models. Sites named in the most answers:

- ravio.com: 33 answers, 74 citations
- comprehensive.io: 25 answers, 29 citations
- salary.com: 17 answers, 31 citations
- g2.com: 14 answers, 19 citations
- payscale.com: 13 answers, 23 citations
- github.com: 12 answers, 14 citations
- peoplemanagingpeople.com: 11 answers, 13 citations
- salarycube.com: 10 answers, 15 citations

Pages named in the most answers:

- https://comprehensive.io/content/best-compensation-benchmarking-data-sources (25 answers)
- https://ravio.com/blog/compensation-benchmarking-companies (16 answers)
- https://ravio.com/blog/the-best-tools-for-salary-benchmarking (16 answers)
- https://salary.com/ (12 answers)
- https://peoplemanagingpeople.com/services/best-compensation-benchmarking-companies (11 answers)
- https://ravio.com/blog/free-salary-data (11 answers)
- https://hrtechsaas.com/blog/best-compensation-benchmarking-tools (9 answers)
- https://pin.com/blog/salary-benchmarking-tools (9 answers)
- https://gitnux.org/service/benchmarking (8 answers)
- https://fitsmallbusiness.com/salary-comparison-tools (7 answers)

## Warned against

- Mercer Total Remuneration Survey: 17 of 34 labels negative. "Providers to avoid or use carefully include: - Mercer, Willis Towers Watson (WTW), and Radford when the budget is tight" (Muse Glimmer 30B, negative prompt)
- Radford: 10 of 21 labels negative. "Providers to avoid or use carefully include: - Mercer, Willis Towers Watson (WTW), and Radford when the budget is tight" (Muse Glimmer 30B, negative prompt)
- Willis Towers Watson: 6 of 13 labels negative. "explicitly advise skipping these providers unless you have a significant budget and need deep integration capabilities" (Muse Glimmer 30B, negative prompt)
- Glassdoor: 6 of 16 labels negative. "Avoid those relying heavily on unverified crowdsourced data (e.g., from Glassdoor or Levels.fyi), as it can be biased or inflated." (Grok 4.20, scale prompt)

## Movement since October 2026

Floor: 13 points of share, measured on the calibration repeat.
New leader: Pave (19%) to Payscale (18%). Margin 3 points, within noise.

- Ravio: 7% to 0%, −7. Inside the floor by 6 points · rank 5 → 29 of 135.
- SalaryCube: 10% to 4%, −6. Inside the floor by 7 points · rank 3 → 6 of 135.
- Salary.com CompAnalyst: 9% to 4%, −5. Inside the floor by 8 points · rank 4 → 5 of 135.
- Pave: 19% to 15%, −4. Inside the floor by 9 points · rank 1 → 2 of 135.
- Figures: 0% to 4%, +4. Inside the floor by 9 points · rank 20 → 7 of 135.
- Glassdoor: 0% to 4%, +4. Inside the floor by 9 points · rank 145 → 8 of 135.
- Radford: 5% to 2%, −3. Inside the floor by 10 points · rank 7 → 27 of 135.
- Comprehensive.io: 7% to 9%, +2. Inside the floor by 11 points · rank 6 → 3 of 135.
- Willis Towers Watson: 0% to 2%, +2. Inside the floor by 11 points · rank 143 → 28 of 135.
- WTW: 2% to 0%, −2. Inside the floor by 11 points · rank 19 → 133 of 135.
- Payscale: 19% to 18%, −1. Inside the floor by 12 points · rank 2 → 1 of 135.
- Mercer Total Remuneration Survey: 3% to 4%, +1. Inside the floor by 12 points · rank 9 of 135, unchanged.

## Record

- Method: https://www.gane.ai/index/hr/methodology
- License: CC BY 4.0. Cite as Gane Index, November 2026 Edition, www.gane.ai/index/hr.