AI BENCHY Compare
Google: Gemini 3.1 Pro Preview vs StepFun: Step 3.5 Flash
Vergelijken:
Benchmarks gegenereerd uit AI BENCHY-testsuites op: 2026-03-06
| Metriek | Google: Gemini 3.1 Pro Preview medium Releasedatum: 2026-02-19 | StepFun: Step 3.5 Flash medium Releasedatum: 2026-02-01 Gratis beschikbaar |
|---|---|---|
| Rang | #2 | #13 |
| Gem. score | 9.4 | 7.4 |
| Consistentie | 10.0 | 9.1 |
| Kosten per resultaat | 3.417 | 0.000 |
| Totale kosten | $0.513 | $0.000 |
| Correcte tests | ||
| Slaagpercentage per poging | 93.8% | 68.8% |
| Instabiele tests | 0 | 2 |
| Totaal runs | 48 (16 x 3) | 48 (16 x 3) |
| Uitvoer-tokens | 1,521 | 71,452 |
| Redeneer-tokens | 35,656 | 155,147 |
| Responstijd (gem.) | 16.60s | 29.10s |
| Responstijd (max) | 40.61s | 170.45s |
| Responstijd (totaal) | 149.36s | 290.96s |
Score vs totale kosten
Responstijd (gem.)
Gem. score vs Responstijd (gem.)
Categorie-uitsplitsing
| Anti-AI-trucs | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 10.0 | 10.0 | 100.0% | 0 | 9.52s | 106 | 2,533 | |
| StepFun: Step 3.5 Flash | 10.0 | 10.0 | 100.0% | 0 | 18.54s | 13,924 | 17,208 |
| Gecombineerd | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 9.0 | 10.0 | 100.0% | 0 | 40.61s | 432 | 9,281 | |
| StepFun: Step 3.5 Flash | 10.0 | 10.0 | 100.0% | 0 | 29.57s | 1,176 | 12,984 |
| Gegevensparsering en extractie | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 9.9 | 10.0 | 100.0% | 0 | 7.72s | 279 | 3,904 | |
| StepFun: Step 3.5 Flash | 10.0 | 10.0 | 100.0% | 0 | 15.01s | 600 | 13,886 |
| Domeinspecifiek | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 7.0 | 10.0 | 66.7% | 0 | 32.73s | 18 | 12,424 | |
| StepFun: Step 3.5 Flash | 4.0 | 7.2 | 44.4% | 1 | 170.45s | 45,350 | 90,436 |
| Algemene intelligentie | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 10.0 | 10.0 | 100.0% | 0 | 11.77s | 108 | 1,179 | |
| StepFun: Step 3.5 Flash | 6.0 | 10.0 | 0.0% | 0 | 6.54s | 2,214 | 2,584 |
| Instructies opvolgen | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 10.0 | 10.0 | 100.0% | 0 | 9.56s | 72 | 2,236 | |
| StepFun: Step 3.5 Flash | 9.0 | 6.8 | 83.3% | 1 | 4.98s | 2,284 | 3,412 |
| Puzzle Solving | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 10.0 | 10.0 | 100.0% | 0 | 7.15s | 232 | 3,117 | |
| StepFun: Step 3.5 Flash | 4.0 | 10.0 | 33.3% | 0 | 7.72s | 5,629 | 10,835 |
| Toolaanroepen | Score | Consistentie | Slaagpercentage per poging | Instabiele tests | Correcte tests | Responstijd (gem.) | Uitvoer-tokens | Redeneer-tokens |
|---|---|---|---|---|---|---|---|---|
| Google: Gemini 3.1 Pro Preview | 10.0 | 10.0 | 100.0% | 0 | 23.15s | 274 | 982 | |
| StepFun: Step 3.5 Flash | 10.0 | 10.0 | 100.0% | 0 | 11.91s | 275 | 3,802 |
Snelle vergelijking
Vergelijkingspaar wisselen
GPT-5.2 ChatnonevsStep 3.5 FlashmediumGratis beschikbaarGemini 3.1 Flash Lite PreviewlowvsStep 3.5 FlashmediumGratis beschikbaarGPT-5.3 ChatnonevsStep 3.5 FlashmediumGratis beschikbaarGemini 3 Flash PreviewnonevsStep 3.5 FlashmediumGratis beschikbaarGemini 3.1 Flash Lite PreviewnonevsStep 3.5 FlashmediumGratis beschikbaarClaude Sonnet 4.6nonevsStep 3.5 FlashmediumGratis beschikbaarGemini 3.1 Flash Lite PreviewhighvsStep 3.5 FlashmediumGratis beschikbaarGemini 3 Flash PreviewlowvsStep 3.5 FlashmediumGratis beschikbaarQwen3.5 Plus 2026-02-15nonevsStep 3.5 FlashmediumGratis beschikbaarStep 3.5 FlashmediumGratis beschikbaarvsGLM 5noneDeepSeek V3.2nonevsStep 3.5 FlashmediumGratis beschikbaarGemini 3.1 Pro PreviewmediumvsGPT-5.2 Chatnone