# AI coding assistants: what AI models recommend, September 2026

Gane Index, September 2026 Edition. Asked as "AI coding assistant" and as "AI pair programming tool", six framings each, to twelve models with search on, on behalf of a mid-market B2B company. Page: https://www.gane.ai/index/it/developer/ai-coding-assistants/2026-09

This is the permanent record of the September 2026 Edition, published September 17, 2026 and superseded since. The current standing: https://www.gane.ai/index/it/developer/ai-coding-assistants/index.md

**Standing:** GitHub Copilot leads with 67% of first choices; verdict clear leader. 48 first choices across the direct, paraphrase, budget and scale prompts.

## First-choice share

| Rank | Product | Share | Negative rate | Labels |
|---|---|---|---|---|
| 1 | [GitHub Copilot](https://www.gane.ai/index/it/vendors/github-copilot/index.md) | 67% | 13% | 70 |
| 2 | [Windsurf](https://www.gane.ai/index/it/vendors/windsurf/index.md) | 10% | 19% | 27 |
| 3 | [Cursor](https://www.gane.ai/index/it/vendors/cursor/index.md) | 4% | 19% | 52 |
| 4 | [Cline](https://www.gane.ai/index/it/vendors/cline/index.md) | 4% | 8% | 12 |
| 5 | [Amazon Q Developer](https://www.gane.ai/index/it/vendors/amazon-q-developer/index.md) | 2% | 22% | 37 |
| 6 | [Aider](https://www.gane.ai/index/it/vendors/aider/index.md) | 2% | 17% | 12 |
| 12 | [Tabnine](https://www.gane.ai/index/it/vendors/tabnine/index.md) | 0% | 5% | 39 |
| 13 | [Claude Code](https://www.gane.ai/index/it/vendors/claude-code/index.md) | 0% | 20% | 30 |

Rank is by share of first choices among all 39 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: Claude, GitHub Copilot
- GPT-5.4 mini: GitHub Copilot; alternatives Amazon Q Developer, Cursor
- Gemini 3.5 Flash: Augment Code; alternatives GitHub Copilot, Tabnine
- Perplexity Sonar: GitHub Copilot; alternatives Cursor, Sourcegraph Cody
- Grok 4.20: GitHub Copilot; alternatives Amazon Q Developer, Cursor, Gemini Code Assist, Tabnine
- Mistral Small: Cursor, GitHub Copilot; alternatives Claude Code
- DeepSeek V4 Flash: GitHub Copilot; alternatives Amazon Q Developer, Cursor, Tabnine
- Llama 4 Maverick: Cursor, GitHub Copilot; alternatives Claude
- Qwen 3.7 Flash: GitHub Copilot; alternatives Cursor
- Kimi K2: GitHub Copilot; alternatives Tabnine, Windsurf
- GLM 4.7 FlashX: GitHub Copilot; alternatives Amazon Q Developer, Cursor, Tabnine
- MiniMax M2.5: GitHub Copilot; alternatives Claude, Claude Code, Cursor

## Sources the answers cite

64 of 72 answers came back with a source list, from 12 of 12 models. Sites named in the most answers:

- axify.io: 27 answers, 27 citations
- dev.to: 24 answers, 29 citations
- augmentcode.com: 18 answers, 32 citations
- memeburn.com: 18 answers, 18 citations
- daily.dev: 17 answers, 19 citations
- zapier.com: 13 answers, 13 citations
- learn.g2.com: 12 answers, 12 citations
- layer3labs.io: 11 answers, 12 citations

Pages named in the most answers:

- https://axify.io/blog/the-best-ai-coding-assistants-a-full-comparison-of-20-tools (27 answers)
- https://memeburn.com/best-ai-coding-tools (18 answers)
- https://learn.g2.com/best-ai-coding-assistants (12 answers)
- https://zapier.com/blog/ai-coding-tools (12 answers)
- https://daily.dev/blog/best-ai-coding-assistants-comparison (11 answers)
- https://pricepertoken.com/coding-assistants (10 answers)
- https://layer3labs.io/guides/best-ai-coding-tools-for-business (9 answers)
- https://tech-insider.org/best-ai-coding-assistants-2026 (9 answers)
- https://forbes.com/councils/forbestechcouncil/2025/03/21/ai-coding-assistants-17-risks-and-how-to-mitigate-them (8 answers)
- https://gartner.com/reviews/market/ai-code-assistants (8 answers)

## Warned against

- Cursor: 10 of 52 labels negative. "avoid standard free-tier or public instances of major AI coding assistants (such as GitHub Copilot standard, Cursor, or Windsurf)" (Qwen 3.7 Flash, negative prompt)
- GitHub Copilot: 9 of 70 labels negative. "You should generally avoid standard free-tier or public instances of major AI coding assistants (such as GitHub Copilot standard, Cursor, or Windsurf)" (Qwen 3.7 Flash, negative prompt)
- Claude Code: 6 of 30 labels negative. "Research found Claude Code silently ignored its own deny rules when commands exceeded 50 subcommands" (Kimi K2, negative prompt)
- Windsurf: 5 of 27 labels negative. "avoid standard free-tier or public instances of major AI coding assistants (such as GitHub Copilot standard, Cursor, or Windsurf)" (Qwen 3.7 Flash, negative prompt)

## Record

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