Google: Gemini 3.5 Flash
15 benchmarks
Leaderboard position
#195
Overall rank (of 1126 models)
1922
Elo rating
15
Benchmarks
Top 17.3%
Percentile
In Google (of 73 models)
Above: Google: gemini-3.5-flash-lite #194
Below: NVIDIA Corporation: nvidia-nemotron-3.5-super-vl #196
Benchmark highlights
Best benchmark
MineBench
1482
Worst benchmark
Agent Arena - Net Improvement (Snapshot from july 12, 2026
-0.9
Nearest competitor from another company
The model from another company with the most similar benchmark scores.
appeared as nearest in 2 of 15 benchmarks
2/15
13.3%
2/15
13.3%
2/15
13.3%
| Model | Votes | % of benchmarks |
|---|---|---|
| MiniMax: MiniMax M3 | 2 | 13.3% |
| Anthropic: Claude Opus 4.7 | 2 | 13.3% |
| DeepSeek: DeepSeek V4 Pro | 2 | 13.3% |
| DeepSeek: deepseek-v3.2 | 1 | 6.7% |
| OpenAI: GPT-5.5 Pro | 1 | 6.7% |
| DeepSeek: deepseek-v4-flash-0731 | 1 | 6.7% |
| Mistral AI: Mistral Large 4 (Le Chonk) | 1 | 6.7% |
| OpenAI: gpt-6.1-sol | 1 | 6.7% |
| inclusionAI: ling-3.1-flash | 1 | 6.7% |
| Qwen: Qwen3.7 Max | 1 | 6.7% |
Benchmarks
MineBench
1482
Text Arena (Snapshot from 20 june 2026)
1476
MMMU-Pro
83.6
SimpleBench
76.7
Artificial Analysis Agentic Index (Maximum Reasoning)
70.3
Finance Agent v2
57.9
Artificial Analysis Intelligence Index v4.0 (Maximum Reasoning)
55.3
Artificial Analysis Intelligence Index v4.1 (Maximum Reasoning)
50
RedlineBench
45.1
DeepSWE 1.1
37
KingBench 3
34.29
AA-Omniscience Index
23
Agent Arena - Confirmed Success (Snapshot from 20 june 2026)
1.06
CowQA-Rapid
0.93
Agent Arena - Net Improvement (Snapshot from july 12, 2026
-0.9
Predicted scores
Estimated from 25 similar models and the global Elo rating. These are projections, not measured results.