Compare Models

Gemini 3.7 Flash

Gemini 3.7 Flash is a model from Google, released August 13, 2026. Its best result is #6 of 151 on LegalBench.

Release Date: Aug 13, 2026

Developer Google  🇺🇸
Context Window 1M
Max Output Tokens 66k
Token Costs (in/out) $1.50/7.50
Weights Private
Input Modalities

Accuracy

56.02 %

Avg. Cost (In/Out)

$ 1.50 / $ 7.50

Latency

228 min 56 s

Vals Index
BenchmarksAccuracyRankings
Vals RSI Index

18.27%

22/27
Code Migration

34.80%

±4.22
38/79
CyberBench v1.1

43.75%

±2.21
43/48
EMB

71.33%

±2.26
12/76
Finance Agent (v2)

59.04%

±0.27
7/81
Legal Research Bench

34.62%

±3.31
52/94
MedCode

53.39%

±2.12
9/107
MedScribe

83.94%

±2.00
40/109
MortgageTax

66.65%

±0.92
33/98
ProofBench v1.1

58.00%

±4.96
25/53
SRE Bench

4.58%

±1.29
18/37
Tax Agent Bench

22.80%

±3.03
51/85
TaxEval v2

74.73%

±0.85
31/145
Vibe Code Bench v1.1

70.39%

±4.84
52/128
GPQA Diamond

93.94%

±1.49
5/138
IOI

67.83%

±3.61
15/46
LiveCodeBench

88.65%

±0.92
5/143
LegalBench

87.26%

±0.42
6/151
MMLU Pro

90.12%

±0.30
6/138
MMMU Pro

88.96%

±0.75
5/93
ProgramBench

0.00%

±0.00
38/66
SkillsBench

65.89%

±4.55
3/35
SWE-bench

80.80%

±1.76
23/88
Terminal-Bench 4.0

12.12%

±3.03
45/64
Terminal-Bench-Science

7.14%

±3.10
22/45
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : Google
Temperature: Default
Top P: Default
Top K: Default
Max Output Tokens: 65,536
Reasoning Effort: high

Compared with

Gemini 3.7 Flash vs DeepSeek V4.1 Flash 14 shared benchmarks Gemini 3.7 Flash vs GPT-6 Luna 12 shared benchmarks Gemini 3.7 Flash vs Ember-1 8 shared benchmarks Gemini 3.7 Flash vs GPT-5.6 Luna 20 shared benchmarks Gemini 3.7 Flash vs Claude Sonnet 5 21 shared benchmarks Gemini 3.7 Flash vs Grok 4.6 20 shared benchmarks Gemini 3.7 Flash vs Kimi K3 20 shared benchmarks Gemini 3.7 Flash vs Hy4 Preview 13 shared benchmarks Gemini 3.7 Flash vs GPT-5.6 Terra 21 shared benchmarks Gemini 3.7 Flash vs Step 5 Preview 8 shared benchmarks