Environmental Impact

Every prompt
tracked & measured

Quantified sustainability gains from prompt optimization, smart routing, and carbon-aware inference. Real numbers from real usage.

Stats update as the framework is used
Global Reach

Growing every day

Real usage statistics from across the GreenInfer platform and GreenPrompts Optimizer combined.

5,847
Total Users
+5 today
38,456
Prompts Optimized
Across all sessions
31.2
Wh Energy Saved
vs always-large baseline
14.8g
CO₂ Avoided
This month
97%
Energy Saved
Small vs Large model
Recent activity
Anonymized usage events across the platform
Model Routing Distribution
Percentage of queries routed to each tier (balanced mode)
Llama 1B (S)
55% ~0.9 mWh
Llama 8B (M)
35% ~3.8 mWh
Llama 70B (L)
10% ~48 mWh
Avg energy per query:4.5 mWh
Baseline (always large):~48 mWh
Where the savings come from
Energy reduction by framework component
✂️ Prompt Optimization-38% tokens, -38% energy
⚡ Model Tier Routing-91% per small-routed query
🌿 Carbon-Aware Scheduling-12% CO₂ from grid timing
🔄 Cascade Confidence GatingPrevents unnecessary escalation
Combined avg reduction~73%
Pareto Frontier - Accuracy vs Energy
Optimal routing curve across model tiers
GreenInfer selects the lowest-energy point that meets accuracy requirements for each prompt.
CO₂ Emissions - GreenInfer vs Baseline
Across 20 benchmark queries
GreenInfer
Always-Large Baseline
GreenInfer total: 0.84g CO₂
Baseline total: 3.12g CO₂
Grid Awareness

ERCOT Carbon Intensity

CURRENT INTENSITY Estimated
198 gCO₂/kWh
0 (Clean)400 (Moderate)800+ (Heavy)
Grid favorable - Balanced routing active
Above 400 gCO₂/kWh: eco mode forced. Below 150: performance unlocked.
BENCHMARK RESULTS (20 PROMPT TEST SET)
Routing accuracy55%
Energy saved vs baseline97%
CO₂ avoided (test set)183.6g
Complexity classifier accuracy98.9%