Google: Gemini 4 Argon
Capabilities
Strengths
- Exceptional reasoning and multi-step problem-solving
- Best-in-class code generation and software engineering
- Multimodal mastery: vision, audio, video understanding
- 1M token context for complex long-document tasks
- Structured output and tool use for agents
Weaknesses
- Slower inference due to extended thinking overhead
- High output pricing ($10/M tokens) limits scale
- Closed-source weights; no self-hosting option
- Limited transparency into training and reasoning processes
- Reasoning depth may be overkill for simple tasks
Pricing
Input / 1M tokens
$2
Output / 1M tokens
$10
Context window
1,048.576k
Transparency
Open weights
0.0 / 10
Open training data
0.0 / 10
Open methodology
3.5 / 10
Licence openness
1.5 / 10
Provider disclosure
5.0 / 10
FMTI company score
N/A
Closed-source flagship model. Google publishes some research and safety practices but withholds weights, training data, and detailed methodology. Extended thinking capability adds reasoning overhead but limits transparency into decision pathways.
Sustainability
Inference energy
N/A
Training footprint
N/A
Provider infrastructure
7.5 / 10
Google operates significant renewable energy infrastructure across data centres, though specific inference energy metrics for Gemini 4 Argon are not publicly disclosed. Larger model size and extended thinking likely increase per-inference carbon cost.