AI lab · 26 models tracked
Google models
Every Google model released in the last 18 months, with release dates, context windows, API prices and benchmark scores. Newest: Gemini 4 Argon, 7d ago.
- Models · 30 days
- 2 of 26
- Latest release
- 7d ago
- Cheapest input / 1M
- $0.09 Gemma 4 26B A4B
- Largest context
- 1M Gemini 4 Argon
Release history
Timeline
Google ships a new model about every 20 days. Each dot is a release; click one to open it.
Lineup
All Google models
Benchmarks
Google's best scores
The top Google model on each benchmark, and where it ranks among all 222 tracked models.
| Benchmark | Best Google model | Score | Overall rank |
|---|---|---|---|
| FrontierSWESoftware engineering | Gemini 4 Argon | 55.0% | #5of 19 |
| DeepSWESoftware engineering | Gemini 3.8 Flash | 73.8% | #2of 26 |
| CursorBenchSoftware engineering | Gemini 4 Argon | 52.9% | #3of 16 |
| FrontierCodeSoftware engineering | Gemini 4 Argon | 47.5% | #11of 41 |
| SWE-bench VerifiedSoftware engineering | Gemini 3.5 Flash | 79.3% | #2of 26 |
| Terminal-BenchSoftware engineering | Gemini 3.1 Pro Preview | 80.2% | #3of 40 |
| MirrorCodeSoftware engineering | Gemini 3.1 Pro Preview | 8.9% | #9of 9 |
| GBAEvalSoftware engineering | Gemini 3.5 Flash | 6.7% | #14of 23 |
| GSOSoftware engineering | Gemini 3.1 Pro Preview | 22.6% | #13of 20 |
| SciCodeSoftware engineering | Gemini 4 Argon | 61.8% | #3of 93 |
| WeirdML v3Software engineering | Gemini 4 Argon | 20.2% | #5of 22 |
| METR time horizonSoftware engineering | Gemini 3.1 Pro Preview | 6.4 h | #2of 13 |
| APEX-AgentsAgents & computer use | Gemini 4 Argon | 82.4% | #1of 51 |
| Remote Labor IndexAgents & computer use | Gemini 3.8 Flash | 5.8% | #7of 13 |
| GDPvalAgents & computer use | Gemini 2.5 Pro | 23.3% | #6of 7 |
| Vending-Bench 2Agents & computer use | Gemini 4 Argon | $13,718 | #3of 61 |
| PostTrainBenchAgents & computer use | Gemini 3.1 Pro Preview | 22.0% | #10of 11 |
| Lech Mazur WritingWriting & content | Gemini 2.5 Pro | 8.38/10 | #4of 13 |
| DeepResearch BenchWriting & content | Gemini 3 Flash Preview | 49.8% | #7of 20 |
| GDP.pdfWriting & content | Gemini 4 Argon | 28.4% | #8of 38 |
| SimpleQA VerifiedWriting & content | Gemini 4 Argon | 76.0% | #1of 66 |
| WebDev ArenaDesign & visual | Gemini 4 Argon | 1680 | #8of 88 |
| Blueprint-Bench 2Design & visual | Gemini 4 Argon | 54.4% | #1of 32 |
| Furniture AssemblyDesign & visual | Gemini 4 Argon | 72.5% | #5of 34 |
| VPCTDesign & visual | Gemini 3 Flash Preview | 72.6% | #2of 9 |
| SpatialViz-BenchDesign & visual | Gemini 2.5 Pro | 44.7% | #1of 3 |
| GeoBenchDesign & visual | Gemini 3 Flash Preview | 88.0% | #1of 3 |
| ExploitBenchCybersecurity | Gemini 3.1 Pro Preview | 26.1% | #3of 8 |
| GPQA DiamondReasoning & knowledge | Gemini 4 Argon | 96.4% | #1of 92 |
| Humanity's Last ExamReasoning & knowledge | Gemini 3.1 Pro Preview | 46.4% | #3of 23 |
| ARC-AGI-2Reasoning & knowledge | Gemini 4 Argon | 91.3% | #5of 63 |
| ARC-AGI-1Reasoning & knowledge | Gemini 4 Argon | 98.5% | #1of 63 |
| SimpleBenchReasoning & knowledge | Gemini 3.1 Pro Preview | 79.6% | #3of 43 |
| CritPtReasoning & knowledge | Gemini 4 Argon | 27.1% | #13of 96 |
| Chess PuzzlesReasoning & knowledge | Gemini 3.7 Flash | 66.0% | #5of 70 |
| LMCAReasoning & knowledge | Gemini 4 Argon | 63.7% | #6of 96 |
| FrontierMathMaths | Gemini 4 Argon | 86.0% | #10of 72 |
| FrontierMath Tier 4Maths | Gemini 4 Argon | 78.0% | #9of 62 |
| ProofBenchMaths | Gemini 4 Argon | 99.0% | #4of 70 |
| OTIS Mock AIMEMaths | Gemini 4 Argon | 100.0% | #1of 90 |
Directory