14 ranked models in a catalog of 28 text models. Gemini 3.7 Flash leads this lab at #12 overall.
Ranked models
Global ranks · current snapshot| Rank | Model | SI Score | coding | math | reasoning | preference | Confidence | Blended price | Context | Released |
|---|---|---|---|---|---|---|---|---|---|---|
| #12 | Gemini 3.7 Flash | 70.0 | 58.9 | 80.7 | 83.9 | 80.7 | 100% | $1.50 | 1M | Aug 13, 2026 |
| #17 | Gemini 3.8 Flash | 68.8 | 56.4 | 77.5 | 74.5 | 81.7 | 100% | $1.50 | 1M | Sep 2, 2026 |
| #19 | Gemini 3.5 Flash | 68.3 | 64.0 | 74.1 | 79.7 | 80.0 | 100% | $3.38 | 1M | May 19, 2026 |
| #22 | Gemini 3.1 Pro Preview | 68.1 | 57.0 | 75.9 | 77.1 | 80.3 | 100% | $4.50 | 1M | Feb 19, 2026 |
| #35 | Gemini 3.6 Flash | 65.5 | 58.6 | 72.7 | 76.2 | 80.1 | 100% | $1.50 | 1M | Jul 21, 2026 |
| #42 | Gemma 4 26B A4B ITOpen weights | 63.4 | — | 82.2 | 73.2 | 76.3 | 64% | $0.00 | 262K | Apr 2, 2026 |
| #45 | Gemma 4 31B ITOpen weights | 63.0 | — | 73.3 | 75.8 | 77.1 | 64% | $0.00 | 262K | Apr 2, 2026 |
| #68 | Gemini 3 Pro Preview | 58.6 | 63.7 | 91.4 | 55.9 | — | 68% | — | 1M | Nov 18, 2025 |
| #89 | Gemini 3.5 Flash Lite | 55.2 | 57.2 | 55.9 | 45.9 | 76.4 | 100% | $0.85 | 1M | Jul 21, 2026 |
| #91 | Gemini 3 Flash Preview | 54.8 | 61.9 | 46.2 | 44.0 | 77.1 | 93% | $1.13 | 1M | Dec 17, 2025 |
| #119 | Gemini 2.5 Pro | 45.2 | 48.5 | 26.7 | 27.4 | 78.4 | 90% | $3.44 | 1M | Jun 5, 2025 |
| #120 | Gemma 3 12B ITOpen weights | 45.1 | — | 16.7 | 39.5 | 66.2 | 64% | — | 131K | Mar 12, 2025 |
| #122 | Gemma 3 27B ITOpen weights | 42.0 | 8.1 | 56.9 | 47.7 | 68.8 | 74% | — | 131K | Mar 12, 2025 |
| #128 | Gemma 3 4B ITOpen weights | 39.2 | — | 7.5 | 23.2 | 61.2 | 64% | — | 131K | Mar 12, 2025 |
Pillars use a 0–100 scale. * marks an imputed neutral prior where the model has no scored results in that pillar; it is not a measured benchmark result. Prices are USD per 1M tokens; see the blend and price comparison.
Scores and releases
Provisional models
14 awaiting rank-eligible evidence- Gemini Flash-Lite Latest
No usable open benchmark results found
- DiffusionGemma 26B A4B IT
No usable open benchmark results found
- Gemma 4 12B IT
No usable open benchmark results found
- Gemini Flash Latest
No usable open benchmark results found
- Gemini 3.1 Flash Lite
Evidence completeness 32.0%; at least 50% required
- Gemma 4 E2B IT
No usable open benchmark results found
- Gemma 4 E4B IT
No usable open benchmark results found
- Gemini 3.1 Flash Lite Preview
Evidence completeness 48.0%; at least 50% required
- Gemini 3.1 Pro Preview Custom Tools
Results cover 1 pillar; at least 2 required
- Gemini 2.5 Computer Use Preview
No usable open benchmark results found
- Gemini 2.0 Flash
Evidence completeness 28.0%; at least 50% required
- Gemini 2.5 Flash-Lite
Results cover 1 pillar; at least 2 required
- Gemini 2.5 Flash
Evidence completeness 47.0%; at least 50% required
- Gemini 2.0 Flash-Lite
No usable open benchmark results found
Latest Google news
All AI news ↗Introducing Playground: Create and play custom games
EmbeddingGemma 2: an open, lightweight multimodal embedding model
The latest AI news we announced in September 2026
Gemini 4 Argon: our next era of frontier intelligence
Introducing SynthID Bio
Watch the winning trailer from the Future Vision XPRIZE, The Gifted.