✍️ AI terbaik untuk generasi teks
Peringkat menggabungkan beberapa skor benchmark yang dibobot sesuai kasus penggunaan ini. Diperbarui Agustus 2026.
Jawaban cepat
1. Claude Mythos 5 (Anthropic, Skor: 76.2), 2. Claude Mythos Preview (Anthropic, Skor: 69.6), 3. Claude Sonnet 4.6 (max) (Anthropic, Skor: 68.8).
| # | Model | Penyedia | Skor | Kecepatan | Harga (output) |
|---|---|---|---|---|---|
| 1 | Claude Mythos 5 | Anthropic | 76.2 | 50 t/s | $50/M |
| 2 | Claude Mythos Preview | Anthropic | 69.6 | 80 t/s | $15/M |
| 3 | Claude Sonnet 4.6 (max) | Anthropic | 68.8 | 50 t/s | $15/M |
| 4 | Gemini 3.1 Pro | 61.1 | 136 t/s | $12/M | |
| 5 | GPT-5.5 Pro | OpenAI | 50.0 | 60 t/s | $10/M |
| 6 | Gemini 3 Pro | 48.0 | 120 t/s | $5/M | |
| 7 | Claude Opus 4 5 Thinking | Anthropic | 48.0 | 50 t/s | $15/M |
| 8 | Deepseek V3 2 Reasoning | DeepSeek | 48.0 | 40 t/s | $1.1/M |
| 9 | Claude 4 5 Sonnet Thinking | Anthropic | 48.0 | 50 t/s | $15/M |
| 10 | Claude 4 1 Opus Thinking | Anthropic | 48.0 | 50 t/s | $15/M |
| 11 | Kimi K2 Thinking | Moonshot AI | 48.0 | 40 t/s | $4/M |
| 12 | O3 Pro | OpenAI | 48.0 | 60 t/s | $10/M |
| 13 | Claude 4 5 Haiku Reasoning | Anthropic | 48.0 | 50 t/s | $15/M |
| 14 | Nova 2 0 Pro Reasoning Medium | Amazon | 48.0 | 40 t/s | $3.2/M |
| 15 | Claude 4 Opus Thinking | Anthropic | 48.0 | 50 t/s | $15/M |