Side-by-side comparison · models
Gemini 2.5 Flash Lite vs Gemini 3 Pro Preview
What we have checked about each, line by line. Every value shows where it came from and when we last confirmed it.
Models
Gemini 2.5 Flash Lite
Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency.
Models
Gemini 3 Pro Preview
Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts.
Checked facts
Line by line
| Measure | Gemini 2.5 Flash Lite | Gemini 3 Pro Preview |
|---|---|---|
| Input price | $0.100 per million tokens | Not checked yet |
| Output price | $0.400 per million tokens | Not checked yet |
| Price on google-ai-studio | $0.0050 per million cached input tokens | Not checked yet |
| Price on google-ai-studio | $0.050 per million input tokens | Not checked yet |
| Price on google-ai-studio | $0.200 per million output tokens | Not checked yet |
| Latest version | google/gemini-2.5-flash-lite | Not checked yet |
| Time to first token (ms) | 310 | Not checked yet |
| Time to first token (ms) on google-ai-studio | 649 | Not checked yet |
| Output speed (tokens a second) | 278.1 | Not checked yet |
| Output speed (tokens a second) on google-ai-studio | 147 | Not checked yet |
| Tokens served on OpenRouter in a day | 37.6B (for 7 Nov 2026) | Not checked yet |
| Tool calls on OpenRouter in a day | 93.9K (for 7 Nov 2026) | Not checked yet |
| Artificial Analysis Intelligence Index | 10.4 | Not checked yet |
| Gpqa diamond | Not checked yet | 92.6% (for 18 Nov 2025) |
| Swe bench verified | Not checked yet | 72.9% (for 18 Nov 2025) |
| Terminal bench | Not checked yet | 69.4% (for 18 Nov 2025) |
| Arc agi 2 | Not checked yet | 31.1% (for 18 Nov 2025) |
| Arc agi | Not checked yet | 75% (for 18 Nov 2025) |
| Balrog | Not checked yet | 58.1% (for 18 Nov 2025) |
| Cl bench | Not checked yet | 15.8% (for 18 Nov 2025) |
| Deepresearch bench | Not checked yet | 46.3% (for 18 Nov 2025) |
| Frontiermath 2025 02 28 private | Not checked yet | 37.6% (for 18 Nov 2025) |
| Frontiermath tier 4 2025 07 01 private | Not checked yet | 18.8% (for 18 Nov 2025) |
| Gdpval | Not checked yet | 40.3% (for 18 Nov 2025) |
| Geobench | Not checked yet | 84% (for 18 Nov 2025) |
| Gso bench | Not checked yet | 18.6% (for 18 Nov 2025) |
| Metr time horizons | Not checked yet | 71% (for 18 Nov 2025) |
| Proofbench | Not checked yet | 20% (for 18 Nov 2025) |
| Remote labor index | Not checked yet | 1.3% (for 18 Nov 2025) |
| Simplebench | Not checked yet | 76.4% (for 18 Nov 2025) |
| Vpct | Not checked yet | 91% (for 18 Nov 2025) |
| Weirdml | Not checked yet | 69.9% (for 18 Nov 2025) |
| Otis mock aime 2024 2025 | Not checked yet | 91.4% (saturated benchmark: the top models all score near the ceiling) (for 18 Nov 2025) |
A line appears only when at least one record has a value we confirmed from a source. Values we have not confirmed are left out, never guessed. Records of different types share fewer lines.
Add a third
Compare with one more record
How this works
Share it, or build your own
This page works for any two or three records in the directory, of any type. The link stays the same whichever order you name them in, so you can share it as it is. Change the records in the address, for example /directories/vs/claude-code-vs-cursor, or start from our editorial comparisons.