# Greenhouse vs Gem: which do AI models recommend for AI recruiting assistants, September 2026

HR AI Recommendation Index, September 2026 Edition, AI recruiting assistants. Two of twelve models named Greenhouse first on the direct prompt; one named Gem. Page: https://hr-ai-index.com/talent/ai-recruiting-assistants/greenhouse-vs-gem/

| | First-choice share | Rank | Negative rate | Labels | Models naming it |
|---|---|---|---|---|---|
| Greenhouse | 5% | #5 of 13 | 12% | 16 | 10 of 12 |
| Gem | 2% | #8 of 13 | 0% | 10 | 8 of 12 |

## The direct prompt, model by model

- DeepSeek V4 Flash: greenhouse first (first choices: Greenhouse) (alternatives: Ashby, Gem, SmartRecruiters, Workable, hireEZ)
- Kimi K2: greenhouse first (first choices: Greenhouse) (alternatives: GoPerfect, Lever, Workable)
- Qwen 3.7 Flash: gem first (first choices: Gem) (alternatives: Ashby, Autonomy Recruit, Greenhouse, TheHireHub.AI)
- Mistral Small: neither first, one named (first choices: Ashby) (alternatives: Greenhouse, TheHireHub.AI, hireEZ)
- GLM 4.7 FlashX: neither first, one named (first choices: TheHireHub.AI) (alternatives: Ashby, Autonomy Recruit, Greenhouse, Workable)
- Claude Haiku 4.5: neither named (first choices: GoPerfect) (alternatives: Juicebox, Workable)
- GPT-5.4 mini: neither named (first choices: SeekOut) (alternatives: Paradox)
- Gemini 3.5 Flash: neither named (first choices: Ashby) (alternatives: Findem, Humanly, Juicebox)
- Perplexity Sonar: neither named (first choices: GoPerfect) (alternatives: Autonomy Recruit, Workable)
- Grok 4.1 Fast: neither named (first choices: Workable) (alternatives: Ashby, hireEZ)
- Llama 4 Maverick: neither named (first choices: Autonomy Recruit) (alternatives: Humanly, Lever, TheHireHub.AI)
- MiniMax M2.5: neither named (first choices: TheHireHub.AI) (alternatives: Ashby, Lever, Workable)

Share is the count of first choices across the direct, paraphrase, budget and scale prompts over all twelve models, for a mid-market B2B company; rank is within the category. Comparisons are drawn for the top eight products in each category. Published under CC BY 4.0; the output is the models' output, and nothing here is a recommendation by the index.