AI BENCHY Compare
DeepSeek: DeepSeek V3.2 vs Google: Gemini 3.1 Flash Lite Preview
Bandingkan:
Benchmark dihasilkan dari suite pengujian AI BENCHY pada: 2026-03-03
| Metrik | DeepSeek: DeepSeek V3.2 none Rilis: 2025-12-01 | Google: Gemini 3.1 Flash Lite Preview medium Rilis: 2026-03-03 |
|---|---|---|
| Peringkat | #39 | #20 |
| Skor Rata-rata | 4.70 | 6.91 |
| Konsistensi | 8.19 | 10.00 |
| Biaya per hasil | 0.132 | 0.392 |
| Total Biaya | $0.007 | $0.032 |
| Tes benar | ||
| Tingkat lulus per percobaan | 47.6% | 57.1% |
| Tes tidak stabil | 3 | 0 |
| Token output | 4,869 | 1,308 |
| Token penalaran | 0 | 16,986 |
Skor vs Total Biaya
Rincian Kategori
| Trik anti-AI | Skor | Konsistensi | Tingkat lulus per percobaan | Tes tidak stabil | Tes benar | Token output | Token penalaran |
|---|---|---|---|---|---|---|---|
| DeepSeek: DeepSeek V3.2 | 1.00 | 9.68 | 0.0% | 0 | 1,411 | 0 | |
| Google: Gemini 3.1 Flash Lite Preview | 9.00 | 9.99 | 66.7% | 0 | 564 | 3,780 |
| Parsing dan ekstraksi data | Skor | Konsistensi | Tingkat lulus per percobaan | Tes tidak stabil | Tes benar | Token output | Token penalaran |
|---|---|---|---|---|---|---|---|
| DeepSeek: DeepSeek V3.2 | 5.38 | 5.81 | 66.7% | 1 | 1,710 | 0 | |
| Google: Gemini 3.1 Flash Lite Preview | 9.88 | 10.00 | 100.0% | 0 | 279 | 2,952 |
| Spesifik domain | Skor | Konsistensi | Tingkat lulus per percobaan | Tes tidak stabil | Tes benar | Token output | Token penalaran |
|---|---|---|---|---|---|---|---|
| DeepSeek: DeepSeek V3.2 | 1.00 | 7.21 | 22.2% | 1 | 24 | 0 | |
| Google: Gemini 3.1 Flash Lite Preview | 1.00 | 10.00 | 0.0% | 0 | 18 | 5,325 |
| Kepatuhan instruksi | Skor | Konsistensi | Tingkat lulus per percobaan | Tes tidak stabil | Tes benar | Token output | Token penalaran |
|---|---|---|---|---|---|---|---|
| DeepSeek: DeepSeek V3.2 | 8.00 | 9.99 | 50.0% | 0 | 66 | 0 | |
| Google: Gemini 3.1 Flash Lite Preview | 8.00 | 9.99 | 50.0% | 0 | 72 | 2,121 |
| Puzzle Solving | Skor | Konsistensi | Tingkat lulus per percobaan | Tes tidak stabil | Tes benar | Token output | Token penalaran |
|---|---|---|---|---|---|---|---|
| DeepSeek: DeepSeek V3.2 | 7.67 | 7.49 | 88.9% | 1 | 1,136 | 0 | |
| Google: Gemini 3.1 Flash Lite Preview | 7.00 | 10.00 | 66.7% | 0 | 141 | 1,896 |
| Pemanggilan alat | Skor | Konsistensi | Tingkat lulus per percobaan | Tes tidak stabil | Tes benar | Token output | Token penalaran |
|---|---|---|---|---|---|---|---|
| DeepSeek: DeepSeek V3.2 | 10.00 | 10.00 | 100.0% | 0 | 522 | 0 | |
| Google: Gemini 3.1 Flash Lite Preview | 10.00 | 10.00 | 100.0% | 0 | 234 | 912 |
Perbandingan Cepat
Ganti Pasangan Perbandingan
Gemini 3.1 Flash Lite PreviewmediumvsGPT-5.3 ChatnoneGemini 3.1 Flash Lite PreviewmediumvsGPT-5.2 ChatnoneDeepSeek V3.2nonevsGLM 4.7 FlashmediumClaude Sonnet 4.6nonevsGemini 3.1 Flash Lite PreviewmediumGemini 3.1 Flash Lite PreviewmediumvsQwen3.5 Plus 2026-02-15noneGemini 3.1 Flash Lite PreviewmediumvsGLM 5noneDeepSeek V3.2nonevsMiniMax M2.5mediumDeepSeek V3.2nonevsgpt-oss-120bmediumTersedia gratisDeepSeek V3.2nonevsGPT-5 MinimediumDeepSeek V3.2nonevsQwen3.5-FlashmediumDeepSeek V3.2nonevsGPT-5 NanomediumDeepSeek V3.2nonevsQwen3 Coder Nextmedium