OpenAI: GPT-6 Luna
OpenAI
Capabilities
Strengths
- Massive 1M token context window for long documents
- Strong vision and multimodal capabilities
- Extended thinking and structured output support
- Excellent cost-efficiency for balanced tier
- Tools and agentic workflow integration
Weaknesses
- Proprietary weights and methodology limit transparency
- Speed score very low (0.08); inference latency likely high
- Limited public sustainability and energy data
- Analyse and reasoning tasks below balanced-tier average
- Closed-source restricts self-hosting and data control
Pricing
Input / 1M tokens
$0.09999999999999999
Output / 1M tokens
$0.5
Context window
1,050k
Transparency
Open weights
0.0 / 10
Open training data
0.0 / 10
Open methodology
2.5 / 10
Licence openness
1.0 / 10
Provider disclosure
3.5 / 10
FMTI company score
N/A
Closed-source model with limited public methodology disclosure. OpenAI publishes high-level capability reports but does not share training data, weights, or detailed architecture. Transparency baseline reflects industry-standard proprietary practices.
Sustainability
Inference energy
N/A
Training footprint
N/A
Provider infrastructure
6.0 / 10
OpenAI operates distributed infrastructure but does not publicly commit to renewable energy targets or publish detailed carbon metrics. Smaller model variant (Luna vs. Sol) likely has lower per-inference energy than flagship, but overall sustainability posture remains opaque.