google/gemini-4-argon
Gemini 4 Argon is Google’s generally available reasoning model for coding, math, science and agentic tasks. The public gemini-4-argon API accepts text, images, video, audio and PDF and outputs text. Its unified input and output context is 1,048,576 tokens; synchronous output defaults to 65,536 tokens, while long decode can reach 1,048,576 within that shared budget. Implicit caching is supported, but explicit caching and Live API are not. Standard introductory pricing is USD 2 input, 0.10 cached input and 10 output per million tokens through November 30, 2026, then USD 4/0.20/20 from December 1. Batch, Flex and Priority have separate documented rates.
- Context window
- 1M
- Input
- Audio · Document · Image · Text · Video
- Output
- Text
- Released
- 2026-09-30
- Max output
- 1M
- Knowledge cutoff
- 2026-03
- Capabilities
- Attachments, Reasoning, Structured, Tool call
No official pricing available.
Performance
Overall capability
71.8/100
- Coding
- 70.7
- Reasoning
- 72.0
- Professional work
- 72.9
- Knowledge & language
- 64.0
| MCP Atlas | Professional work | Scale AI | 2026-11-10 | |
| Arena Agent Overall | Professional work | LMArena | 2026-11-08 | |
| Harvey Legal Agent Benchmark | Professional work | Vals AI / Harvey | 2026-11-05 | |
| τ³-Banking | Professional work | Mercor | 2026-11-05 | |
| FrontierMath v2 Tiers 1–3 | Reasoning | Epoch AI | 2026-11-05 | |
| BullshitBench v2 | Knowledge & language | BullshitBench | 2026-11-05 | |
| Mystery Game Puzzles | Reasoning | Epoch AI | 2026-11-05 | |
| Chess Puzzles | Reasoning | Epoch AI | 2026-11-05 |
Providers
No providers found
ModelRegistry · Catalog updated 2026-11-10