Best AI model for trades and services, tested on 23 September 2026
The tasks in the AIGROW small-business benchmark that a trade or service business would hand to AI, and how each of 36 models did on every one, marked pass or fail against written checks.
A pilot, 1 run a model on each task. This test ran on 23 September 2026. Since then makers have replaced 16 of its 36 models with newer ones, and it did not include 16 current models on our list. The LLM leaderboard ranks current models on measured results.
- Kimi K3
- 6 of 6
- Gemini 3.1 Pro (preview)
- 6 of 6
- GLM-5.3
- 6 of 6
- Gemini 3.8 Flash
- 6 of 6
- Grok 4.7
- 6 of 6
- Grok 4.3
- 6 of 6
The AIGROW small-business benchmark has 6 tasks a trade or service business would hand to AI. On 23 September 2026, 8 models passed all 6; of those, gpt-oss-120b gave the cheapest passes across the benchmark, $0.422 per thousand. Each task page shows the traps and every model’s mistakes.
Every model on these tasks
Most passes first; a tie goes to the better record across all 20 tasks. The cost of 1,000 passes is the whole benchmark’s, failed runs included.
| Model | Pricing a job from a price list | Writing quote emails | Scoring sales leads | Routing customer messages | Calculating overtime pay | Answering staff handbook questions | Passed | 1,000 passes |
|---|---|---|---|---|---|---|---|---|
| Kimi K3Replaced by Kimi K3.1 Pro | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $18.44 |
| Gemini 3.1 Pro (preview)Replaced by Gemini 4 Argon | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $29.67 |
| GLM-5.3 | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $7.02 |
| Gemini 3.8 Flash | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $7.17 |
| Grok 4.7 | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $9.41 |
| Grok 4.3 | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $4.10 |
| Qwen3.8 27B | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $7.30 |
| gpt-oss-120b | Pass | Pass | Pass | Pass | Pass | Pass | 6 of 6 | $0.422 |
| Qwen3.8 Flash | Fail | Pass | Pass | Pass | Pass | Pass | 5 of 6 | $1.09 |
| GLM-5.3-Flash | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $0.831 |
| DeepSeek V4.1 Flash | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $1.05 |
| DeepSeek V4 Pro | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $5.14 |
| Qwen3.7 Plus | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $3.58 |
| GPT-5.6 SolReplaced by GPT-6.1 Sol | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $11.33 |
| Qwen3.8 Max | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $16.32 |
| GPT-6 Astra | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $27.91 |
| Muse Glimmer 30B | Pass | Fail | Pass | Pass | Pass | Pass | 5 of 6 | $3.10 |
| Claude Sonnet 5Replaced by Claude Sonnet 5.5 | Fail | Fail | Pass | Pass | Pass | Pass | 4 of 6 | $6.70 |
| GPT-5.5Replaced by GPT-6 Astra | Pass | Fail | Pass | Pass | Fail | Pass | 4 of 6 | $17.02 |
| Claude Fable 5.1Replaced by Claude Fable 5.2 | Pass | Fail | Pass | Pass | Fail | Pass | 4 of 6 | $33.11 |
| GPT-5.6 LunaReplaced by GPT-6 Luna | Pass | Fail | Pass | Pass | Fail | Pass | 4 of 6 | $0.677 |
| MiniMax M3 | Fail | Pass | Pass | Pass | Pass | Fail | 4 of 6 | $1.82 |
| Claude Opus 5Replaced by Claude Opus 5.5 | Fail | Fail | Pass | Pass | Fail | Pass | 3 of 6 | $16.89 |
| GPT-5.6 TerraReplaced by GPT-6.1 Terra | Fail | Fail | Pass | Pass | Fail | Pass | 3 of 6 | $6.68 |
| Claude Sonnet 4.6Replaced by Claude Sonnet 5.5 | Fail | Fail | Pass | Pass | Fail | Pass | 3 of 6 | $11.24 |
| Kimi K2.6Replaced by Kimi K3.1 Pro | Fail | Pass | Fail | Pass | Fail | Pass | 3 of 6 | $11.69 |
| Gemma 4 31B | Pass | Fail | Pass | Pass | Fail | Fail | 3 of 6 | $0.558 |
| Gemini 3.1 Flash-LiteReplaced by Gemini 3.5 Flash-Lite | Pass | Pass | Fail | Pass | Fail | Fail | 3 of 6 | $1.41 |
| Llama 4 Maverick | Fail | Pass | Fail | Pass | Fail | Fail | 2 of 6 | $0.665 |
| Mistral Medium 3.5Replaced by Mistral Large 4 | Fail | Pass | Fail | Pass | Fail | Fail | 2 of 6 | $7.40 |
| Gemini 3.5 Flash-Lite | Fail | Pass | Fail | Pass | Fail | Fail | 2 of 6 | $2.47 |
| Claude Haiku 4.5Replaced by Claude Haiku 5.5 | Fail | Fail | Fail | Pass | Fail | Fail | 1 of 6 | $6.62 |
| GPT-5.4 nanoReplaced by GPT-6 Luna | Fail | Pass | Fail | Fail | Fail | Fail | 1 of 6 | $1.51 |
| Mistral Small 4 | Fail | Fail | Fail | Pass | Fail | Fail | 1 of 6 | $1.33 |
| GPT-5.4 miniReplaced by GPT-6.1 Terra | Fail | Fail | Fail | Fail | Fail | Fail | 0 of 6 | $3.35 |
| Ministral 3 14B | Fail | Fail | Fail | Fail | Fail | Fail | 0 of 6 | $1.07 |
The tasks
Each page shows the task as the models saw it, the traps in it, every check in words and what each model got wrong.
- Price a decorating job from survey measurements and the price list
- English. 21 of 36 runs passed.
- Email a garden landscaping quote that handles a VAT request and a protected tree
- English. 16 of 36 runs passed.
- Score five inbound cleaning leads against the sales team's rules
- English. 26 of 36 runs passed.
- Route ten tenant messages to the right team with the right priority
- English. 33 of 36 runs passed.
- Work out a cleaner's weekly gross pay with breaks, overtime and Sunday rates
- English. 19 of 36 runs passed.
- Answer a part-time employee's leave questions from the staff handbook
- English. 25 of 36 runs passed.
Hand this work to an agent
An AIGROW agent does work like this inside your own tools, against rules you approve in writing, and escalates what the rules do not cover instead of guessing.
October 2026 edition