Compare
Up to four models or companies, side by side. Every cell is a number we checked, linked to its source.
| gemini-4-argonGoogle | |
|---|---|
| Price, per million tokens | |
| In / out | $4 / $20 |
| In / out, prompts over 100,000 | not checked yet |
| Cache read | not checked yet |
| Capacity | |
| Context window, tokens | not checked yet |
| Benchmarks | |
| ERQA (Embodied Reasoning), pure reasoning maker's own | 87% |
| Gray Swan Indirect Prompt Injection, attack success across 15 adaptive attempts (lower is better) maker's own | 0.7% |
| shortcut or cheat attempts on coding benchmarks (lower is better) maker's own | 2% |
| LVBench maker's own | 91.7% |
| AutomationBench maker's own | 51.3% |
| DeepSWE v1.1 maker's own | 77.9% |
| Measured by others | not checked yet |
| Availability | |
| Released | not checked yet |
| Where you can buy it | not checked yet |
| Retirement date | not checked yet |
| Open weights | not checked yet |
A dot means we have not checked it yet. The registrar fills "where you can buy it", retirement dates and licences from pages she opens and files.
How much we covered
Stories per week · gemini-4-argon28 Sep to 4 Oct
15 Oct to 11 Oct
012 Oct to 18 Oct
319 Oct to 25 Oct
026 Oct to 1 Nov
02 Nov to 8 Nov
09 Nov to today
0gemini-4-argon. Counts cover only the stories we published, not all AI news.