メインコンテンツにスキップ

Model Providers

The foundation on which the entire ecosystem is built.

Models Experience & Sentiment

While ChatGPT is still the most-used model family, this year Claude is the one that garners the most positive sentiment–a preview of a theme that will echo throughout these survey results.

On the other side of the spectrum Grok is the most-hated model, with many commenters opposing it on moral grounds alone due to its connection with Elon Musk and right-wing ideology.

グループ化:

ソート基準:

1
[ja-JP] tools.models
88.4%
8.6%
7,197
2
[ja-JP] tools.models
82.1%
17%
7,199
3
[ja-JP] tools.models
72.6%
26.4%
7,193
4
[ja-JP] tools.models
31%
54.8%
13.7%
7,182
5
[ja-JP] tools.models
30.9%
63.8%
5%
7,178

経験

  • 使ったことがある: Respondents who have used an item.
  • 知っている: Respondents who have heard about an item, but haven't used it.
  • 知らない: Respondents who have never heard about an item.

所感

  • 肯定的: その技術について詳しく知りたい、あるいは今後も継続して利用したいと考えている回答者。
  • 中立: その技術に対して、肯定・否定のどちらの感情も示さなかった回答者
  • 否定的: その技術についてこれ以上学ぶ意欲がない、あるいは、実際に使用して否定的な印象を抱いた回答者。

Models Ratios Over Time

ChatGPT is still the most-used model, but both Claude and Gemini have seen huge year-over-year progress.

Looking at satisfaction confirms the trend: Claude maintains its top position, and while ChatGPT is still second for now it's losing points at a worrying rate, while Gemini stands out as the one model that's actually gained percentage points since last year.

ChatGPT
Claude
Google Gemini
Llama
Perplexity
Mistral
DeepSeek
Grok
Qwen

モード:

表示:

利用率:

Proportion of respondents having used an item
0%
20%
40%
60%
80%
100%
2025
2026
0%
20%
40%
60%
80%
100%
Most other models cited here were not even part of last year's survey, and it's notable that the top three (Kimi, GLM, and MiniMax) are all being developed by companies based in China,indicating that the future of AI might not belong exclusively to Silicon Valley after all.
What other models are you currently using?
Freeform
Multiple
0%
20%
40%
60%
80%
100%
02
+1
-3
0%
20%
40%
60%
80%
100%
回答数に占める割合(%)
このカテゴリーで回答者が使用した項目数です。
0%
5%
9%
14%
19%
23%
01
0
336
02
1
350
03
2
946
04
3
1,690
05
4
1,305
06
5
966
07
6
681
08
7
462
09
8
294
10
9
228
0%
5%
9%
14%
19%
23%
%の質問回答数
Having the highest amount of positive sentiment was already a great win for Claude, but the fact that it also tops the list of the models people are actually paying for means Anthropic is definitely doing something right.
Out of the models mentioned here, which one do you or your company actually pay for?
Multiple
0%
20%
40%
60%
80%
100%
4,592
3,261
04
🚫 該当なし
1,115
168
110
0%
20%
40%
60%
80%
100%
回答数に占める割合(%)

おすすめのリソース

Practical Prompt Engineering
Sabrina Goldfarb
GitHub

Practical Prompt Engineering

Learn prompting techniques that work consistently across Claude, ChatGPT, Copilot, and Cursor. Transform vague project ideas into structure, actionable development plans.
Craft a Professional AI Developer Setup
Steve Kinney
Temporal

Craft a Professional AI Developer Setup

Boost your coding productivity with AI-assisted tools and agents. Use Cursor and Claude Code for precise inline edits or to delegate larger tasks to background agents.
パートナーのサポートに感謝しています! 詳細をみる。