Gemini 3.5 Flash-Lite
Google DeepMind · released
Google's budget tier: a very fast, low-cost multimodal model for sub-agents, document processing and other high-throughput jobs. It is the fastest model tracked here by a wide margin.
Best for: Highest throughput · Document processing · Sub-agents
- Intelligence Index
- 22.2
- #32 of 38
- Price per 1M tokens
- $0.30 in $2.5 out
- Blended $0.85 at 3:1
- Context window
- 1M
- Up to 66K output
- Output speed
- 346 tokens/s
- $0.19 measured cost per task
Capability profile
One 0–100 score per area, placing this model between the weakest and strongest tracked model across that area's benchmarks.
- Overall37 · #28 of 38
Based on 2 of 3 benchmarks
- Coding14 · #30 of 39
Based on 3 of 5 benchmarks
- Agents & knowledge work36 · #30 of 40
Based on 1 of 6 benchmarks
- Reasoning & science9 · #35 of 39
Based on 2 of 2 benchmarks
- Knowledge & reliability62 · #23 of 39
Based on 2 of 2 benchmarks
- Vision69 · #15 of 31
Based on 2 of 2 benchmarks
Benchmark results
All third-party results on record. Rank is among the tracked models that have been tested on that benchmark.
Overall
#32 of 38· Artificial Analysis
- Vals IndexNot tested
- Arena · Text1455 ±4
#19 of 33· 35,763 votes· Arena
Coding
#27 of 38· Artificial Analysis
- SciCode41.3%
#33 of 38· Artificial Analysis
- Arena · WebDev1440 ±41
#29 of 34· 242 votes· Arena
- Vibe Code BenchNot tested
- Code MigrationNot tested
Agents & knowledge work
- GDPval-AA24.3
#31 of 39· Artificial Analysis
- Arena · AgentNot tested
- Finance Agent v2Not tested
- Excel Modeling BenchmarkNot tested
- Legal Research BenchNot tested
- Tax Agent BenchNot tested
Reasoning & science
- Humanity's Last Exam18.8%
#35 of 39· Artificial Analysis
- CritPt0.0%
#38 of 39· Artificial Analysis
- ARC-AGI-2saturatingNot tested
- ARC-AGI-3Not tested
- GPQA Diamondsaturating83.8%
#21 of 27· Artificial Analysis
Knowledge & reliability
#18 of 39· Artificial Analysis
- AA-LCR76.0%
#31 of 39· Artificial Analysis
Vision
- MMMU-Pro79.0%
#12 of 27· Artificial Analysis
- Arena · Vision1264 ±8
#11 of 19· 7,579 votes· Arena
Reported by Google DeepMind
Numbers the lab published at launch, run on its own harness and settings. Useful context, but not independently verified and never used in the rankings on this site.
- SWE-bench Pro54.2%
- OSWorld-Verified74%
Specifications
- Released
- Jul 21, 2026
- Availability
- Generally available
- Context window
- 1,048,576 tokens
- Max output
- 65,536 tokens
- Accepts
- Text, Images, Audio, Video, PDFs
- Knowledge cutoff
- Mar 2026
- Weights
- Closed
- Reasoning control
- Thinking level minimal to high
- API model ID
gemini-3.5-flash-lite
Pricing
- Input / 1M tokens
- $0.30
- Cached input / 1M
- $0.03
- Output / 1M tokens
- $2.5
- Blended (3:1)
- $0.85
- Measured cost / task
- $0.19
One input rate covers text, image, video and audio. Batch is half price.
Official sources
Closest alternatives
The models nearest to this one on overall score.
23 benchmarks tracked. See methodology and sources.