Gemini 3.5 Flash-LitevsGemini 3.1 Flash-Lite
Across 8 shared benchmarks, Gemini 3.5 Flash-Lite leads overall: Gemini 3.5 Flash-Lite wins 7, Gemini 3.1 Flash-Lite wins 1, with 0 ties and an average score difference of +5.20.
Google Deep Mind · 2026-07-21 · Multimodal model
Google Deep Mind · 2026-05-07 · Multimodal model
Benchmark scores
Grouped by capability, sorted by largest gap within each. 8 shared benchmarks.
AI Agent - Tool Usage
Gemini 3.5 Flash-Lite 2/2| Benchmark | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite | Diff |
|---|---|---|---|
| Terminal-Bench 2.1 | 54123 / 192Thinking (With Tools) | 31.10154 / 192Thinking (With Tools) | +22.90 |
| Terminal-Bench 4.0 | 174 / 86Thinking (With Tools) | 0.5081 / 86Thinking (With Tools) | +0.50 |
Multimodal Understanding
Gemini 3.5 Flash-Lite 2/2| Benchmark | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite | Diff |
|---|---|---|---|
| GDP.pdf | 13.6061 / 118Thinking (No Tools) | 7.8088 / 118Thinking (No Tools) | +5.80 |
| MMMU-Pro | 7951 / 227Thinking (No Tools) | 75.5082 / 227Thinking (No Tools) | +3.50 |
Agent Level Benchmark
Gemini 3.5 Flash-Lite 1/1| Benchmark | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite | Diff |
|---|---|---|---|
| τ³-Banking | 17.5098 / 164Thinking (With Tools) | 9.70131 / 164Thinking (With Tools) | +7.80 |
Coding and Software Engineer
Gemini 3.1 Flash-Lite 1/1| Benchmark | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite | Diff |
|---|---|---|---|
| SciCode | 41.30100 / 130Thinking (No Tools) | 43.4095 / 130Thinking (No Tools) | -2.10 |
General Evaluation
Gemini 3.5 Flash-Lite 1/1| Benchmark | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite | Diff |
|---|---|---|---|
| GPQA Diamond | 83.80181 / 462Thinking (No Tools) | 82.20205 / 462Thinking (No Tools) | +1.60 |
General Knowledge
Gemini 3.5 Flash-Lite 1/1| Benchmark | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite | Diff |
|---|---|---|---|
| HLE | 18.80304 / 563Thinking (No Tools) · Text only | 17.20319 / 563Thinking (No Tools) · Text only | +1.60 |
Specs
| Field | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite |
|---|---|---|
| Publisher | Google Deep Mind | Google Deep Mind |
| Release date | 2026-07-21 | 2026-05-07 |
| Model type | Multimodal model | Multimodal model |
| Architecture | Dense | Dense |
| Parameters | Not available | Not available |
| Context length | 1M | 1M |
| Max output | 64K | 64K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Gemini 3.5 Flash-Lite | Gemini 3.1 Flash-Lite |
|---|---|---|
| Text input | $0.3 / 1M tokens | Not public |
| Text output | $2.5 / 1M tokens | Not public |
| Cache read | $0.03 / 1M tokens | Not public |
One or both models have incomplete public pricing.
Summary
- Gemini 3.5 Flash-Liteleads in:AI Agent - Tool Usage (2/2), Multimodal Understanding (2/2), Agent Level Benchmark (1/1), General Evaluation (1/1), General Knowledge (1/1)
- Gemini 3.1 Flash-Liteleads in:Coding and Software Engineer (1/1)
On average across the 8 shared benchmarks, Gemini 3.5 Flash-Lite scores 5.20 higher.
Largest single-benchmark gap: Terminal-Bench 2.1 — Gemini 3.5 Flash-Lite 54 vs Gemini 3.1 Flash-Lite 31.10 (+22.90).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.