#8 Gemini 3.5 Flash
low- Cost
- $0.068
- Time
- 39.1s
- Tokens
- 7,588 tok
AI BENCHY Compare
Google: Gemini 3.5 Flash vs OpenAI: GPT-5.4
Summary
Gemini 3.5 Flash (low) leads on average score with 9.2 vs 8.5. Gemini 3.5 Flash (low) has the lower benchmark cost at $0.349 vs $1.210. Gemini 3.5 Flash (low) is faster at 3.27s vs 22.35s, with pass rates of 90.5% vs 76.2%.
Recommended modelGemini 3.5 Flash (low)It has the best score here (9.2), while costing about 3.5x less than GPT-5.4 (medium).
Last updated at: 2026-07-14
| Metric | Gemini 3.5 Flash Gemini 3.5 Flash low | GPT-5.4 GPT-5.4 medium |
|---|---|---|
| Score | 9.2 | 8.5 |
| Rank | #8 | #20 |
| Reliability | 10.0 | 10.0 |
| Consistency | 10.0 | 8.6 |
| Tests Correct | ||
| Attempt pass rate | 90.5% | 76.2% |
| Flaky tests | 0 | 4 |
| Total Runs | 63 | 63 |
| Cost per result | 1.834 | 8.640 |
| Total Cost | $0.349 | $1.210 |
| Input Price | $1.500 / 1M | $2.500 / 1M |
| Output Price | $9.000 / 1M | $15.000 / 1M |
| Total Input Tokens | 36,938 | 34,108 |
| Output Tokens | 2,033 | 2,242 |
| Reasoning Tokens | 30,519 | 72,707 |
| Response Time (avg) | 3.27s | 22.35s |
| Response Time (max) | 9.05s | 100.41s |
| Response Time (total) | 68.65s | 469.29s |
Generation showcase
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
#20 GPT-5.4
medium- Cost
- $0.214
- Time
- 199.6s
- Tokens
- 14,349 tok
Top Models by Score
Score vs Total Cost
Response Time (avg)
Score vs Response Time (avg)
Total Output Tokens
Score vs Total Output Tokens
Category Breakdown
| Coding | Score | Consistency | Attempt pass rate | Flaky tests | Tests Correct | Response Time (avg) | Input Tokens | Output Tokens | Reasoning Tokens |
|---|---|---|---|---|---|---|---|---|---|
| Gemini 3.5 Flash | 7.8 | 10.0 | 66.7% | 0 | 6.71s | 8,118 | 458 | 13,420 | |
| GPT-5.4 | 8.8 | 7.8 | 88.9% | 1 | 44.36s | 7,305 | 433 | 24,216 |
Quick Compare
Switch Comparison Pair
Claude Fable 5mediumvsGemini 3.5 FlashlowGemini 3.5 FlashlowvsGPT-5.6 SolhighGemini 3.5 FlashlowvsGPT-5.6 SolmediumGemini 3.5 FlashlowvsQwen3.7 MaxmediumGemini 3.5 FlashlowvsGPT-5.5mediumDeepSeek V4 FlashhighvsGPT-5.4mediumGemini 3.5 FlashlowvsGPT-5.3-CodexmediumGemini 3.5 FlashlowvsQwen3.6 Max PreviewmediumClaude Opus 4.8mediumvsGemini 3.5 FlashlowGPT-5.4mediumvsGLM 5.2highClaude Opus 4.7mediumvsGemini 3.5 FlashlowKAT-Coder-Pro V2.5lowvsGPT-5.4medium