Gemini 2.5 Flash Preview (9/25) (Thinking)

Gemini 2.5 Flash Preview (9/25) (Thinking) is a model from Google, released September 25, 2025. Its best result is #67 of 151 on LegalBench.

Release Date: Sep 25, 2025

Developer Google  🇺🇸
Context Window 1M
Max Output Tokens 66k
Token Costs (in/out) $0.30/2.50
Weights Private
Input Modalities

Accuracy

67.43 %

Avg. Cost (In/Out)

$ 0.30 / $ 2.50

Latency

1 min 1 s

Vals Index
BenchmarksAccuracyRankings

40.33%

±1.92
67/107

78.50%

±1.99
65/109

61.92%

±0.96
58/98

72.40%

±0.87
64/145

76.52%

±2.15
79/138

76.21%

±1.14
79/143

82.63%

±0.41
67/151

83.66%

±0.36
72/138

80.75%

±0.95
43/93

21.35%

±4.37
57/67
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : Google
Temperature: 1
Top P: Default
Top K: Default
Max Output Tokens: 65,535

Updates

Sep 26, 2025

We evaluated Gemini 2.5 Flash Preview (9/25) (Thinking) (and also the Flash Lite model) and found the following:

  • Compared to the previous version, the update shows improvements on GPQA Diamond, with a ~17% increase over the previous version.

  • Flash improved on Terminal-Bench (+5%), GPQA Diamond (+17.2%), and our private Corp Fin Benchmark (+4.4%). It also ranks #3/38 on MMMU Pro and #6/20 on SWE-bench Verified (for the thinking model), while delivering performance at half the cost of similar models.

  • Flash Lite matches Flash on several public benchmarks, making it a very cost-effective option. However, Flash outperforms Lite by ~10% on our private benchmarks (CaseLaw v2, TaxEval, Mortgage Tax).

  • Flash delivers competitive performance at a fraction of the cost of other foundation models.

Overall, the latest update to Gemini 2.5 Flash is a highly efficient model that balances strong performance with low cost.