ModelPilot Logo
Model Environmental Impact Estimates

AI Model Environmental Leaderboard

Compare the environmental impact of popular AI models. Estimated CO₂e emissions, energy consumption, and overall rankings based on our calibrated environmental model.

Environmental Performance Rankings
Models ranked by overall environmental impact (lower emissions = better ranking)
RankModelProviderParametersCO₂e per 1M Tokens
Input / Output
Energy per 1M Tokens
Input / Output
Cost per 1M Tokens
Input / Output
#1
moonshotai/kimi-vl-a3b-thinking
moonshotai
0.75B active (3B total)
~62.32 µg (±25%)🟢
31.16 µg/93.49 µg
~119.9mWh (±25%)
59.9mWh/179.8mWh
$0.0000/$0.0000
Avg: $0.0000
#2
anthropic/claude-3.5-haiku
anthropic
13B
Est.
~136 µg (±25%)🟢
67.99 µg/204 µg
~523.0mWh (±25%)
261.5mWh/784.5mWh
$0.000001/$0.000004
Avg: $0.000002
#3
meta-llama/llama-4-maverick
meta-llama
4.25B active (17B total)
~137.5 µg (±25%)🟢
68.73 µg/206.2 µg
~624.8mWh (±25%)
312.4mWh/937.3mWh
$0.0000/$0.000001
Avg: $0.0000
#4
meta-llama/llama-4-scout
meta-llama
4.25B active (17B total)
~137.5 µg (±25%)🟢
68.73 µg/206.2 µg
~624.8mWh (±25%)
312.4mWh/937.3mWh
$0.0000/$0.0000
Avg: $0.0000
#5
meta-llama/llama-3.1-8b-instruct
meta-llama
8B
~226.4 µg (±25%)🟢
113.2 µg/339.6 µg
~1.0Wh (±25%)
514.6mWh/1.5Wh
$0.0000/$0.0000
Avg: $0.0000
#6
z-ai/glm-4.5
z-ai
7B
Est.
~234.6 µg (±25%)🟢
117.3 µg/351.8 µg
~521.3mWh (±25%)
260.6mWh/781.9mWh
$0.0000/$0.000001
Avg: $0.000001
#7
moonshotai/kimi-k2
moonshotai
7B
Est.
~249.3 µg (±25%)🟢
124.6 µg/373.9 µg
~479.4mWh (±25%)
239.7mWh/719.1mWh
$0.0000/$0.000002
Avg: $0.000001
#8
anthropic/claude-3.7-sonnet
anthropic
30B
Est.
~340 µg (±25%)🟢
170 µg/509.9 µg
~1.3Wh (±25%)
653.8mWh/2.0Wh
$0.000003/$0.000015
Avg: $0.000009
#9
anthropic/claude-opus-4.1
anthropic
175B
Est.
~340 µg (±25%)🟢
170 µg/509.9 µg
~1.3Wh (±25%)
653.8mWh/2.0Wh
$0.000015/$0.000075
Avg: $0.000045
#10
openai/o3-mini
openai
13B
Est.
~360 µg (±25%)🟢
180 µg/540 µg
~1.3Wh (±25%)
642.9mWh/1.9Wh
$0.000001/$0.000004
Avg: $0.000003
#11
openai/o3-pro
openai
175B
Est.
~360 µg (±25%)🟢
180 µg/540 µg
~1.3Wh (±25%)
642.9mWh/1.9Wh
$0.00002/$0.00008
Avg: $0.00005
#12
qwen/qwen3-30b-a3b-instruct-2507
qwen
7.5B active (30B total)
~623.2 µg (±25%)🟢
311.6 µg/934.9 µg
~1.2Wh (±25%)
599.3mWh/1.8Wh
$0.0000/$0.0000
Avg: $0.0000
#13
x-ai/grok-4
x-ai
30B
Est.
~733 µg (±25%)🟢
366.5 µg/1.10 mg
~1.6Wh (±25%)
814.5mWh/2.4Wh
$0.000003/$0.000015
Avg: $0.000009
#14
openai/gpt-5-nano
openai
8B
Est.
~960 µg (±25%)🟢
480 µg/1.44 mg
~3.4Wh (±25%)
1.7Wh/5.1Wh
$0.0000/$0.0000
Avg: $0.0000
#15
mistralai/ministral-8b
mistralai
8B
~1.955 mg (±25%)🟢
977.4 µg/2.932 mg
~4.3Wh (±25%)
2.2Wh/6.5Wh
$0.0000/$0.0000
Avg: $0.0000
#16
meta-llama/llama-guard-4-12b
meta-llama
12B
~39.93 mg (±25%)🟡
19.96 mg/59.89 mg
~181.5Wh (±25%)
90.7Wh/272.2Wh
$0.0000/$0.0000
Avg: $0.0000
#17
google/gemini-2.0-flash-001
google
300B
Est.
~96.15 mg (±25%)🟡
48.07 mg/144.2 mg
~801.2Wh (±25%)
400.6Wh/1201.8Wh
$0.0000/$0.0000
Avg: $0.0000
#18
mistralai/mistral-small-3.2-24b-instruct
mistralai
24B
~135.1 mg (±25%)🟠
67.56 mg/202.7 mg
~300.3Wh (±25%)
150.1Wh/450.4Wh
$0.0000/$0.0000
Avg: $0.0000
#19
meta-llama/llama-3.3-70b-instruct
meta-llama
70B
~163 mg (±25%)🟠
81.52 mg/244.6 mg
~741.1Wh (±25%)
370.5Wh/1111.6Wh
$0.0000/$0.0000
Avg: $0.0000
#20
mistralai/codestral-2508
mistralai
22B
Est.
~176.9 mg (±25%)🟠
88.47 mg/265.4 mg
~393.2Wh (±25%)
196.6Wh/589.8Wh
$0.0000/$0.000001
Avg: $0.000001
#21
openai/gpt-5-mini
openai
20B
Est.
~197.5 mg (±25%)🟠
98.75 mg/296.3 mg
~705.4Wh (±25%)
352.7Wh/1058.1Wh
$0.0000/$0.000002
Avg: $0.000001
#22
deepseek/deepseek-chat-v3.1
deepseek
37B active (671B total)
~253 mg (±25%)🟠
126.5 mg/379.5 mg
~486.6Wh (±25%)
243.3Wh/729.9Wh
$0.0000/$0.000001
Avg: $0.000001
#23
z-ai/glm-4.5-air
z-ai
100B
Est.
~266.7 mg (±25%)🟠
133.3 mg/400 mg
~592.6Wh (±25%)
296.3Wh/888.9Wh
$0.0000/$0.000001
Avg: $0.000001
#24
openai/gpt-5-chat
openai
175B
Est.
~286.5 mg (±25%)🟠
143.2 mg/429.7 mg
~1023.2Wh (±25%)
511.6Wh/1534.8Wh
$0.000001/$0.00001
Avg: $0.000006
#25
qwen/qwen3-coder
qwen
120B active (480B total)
~340.1 mg (±25%)🟠
170.1 mg/510.2 mg
~654.1Wh (±25%)
327.0Wh/981.1Wh
$0.0000/$0.000001
Avg: $0.000001
#26
google/gemini-2.5-flash
google
300B
Est.
~343.4 mg (±25%)🟠
171.7 mg/515.1 mg
~2861.5Wh (±25%)
1430.8Wh/4292.3Wh
$0.0000/$0.000003
Avg: $0.000001
#27
google/gemini-2.5-pro
google
300B
Est.
~343.4 mg (±25%)🟠
171.7 mg/515.1 mg
~2861.5Wh (±25%)
1430.8Wh/4292.3Wh
$0.000001/$0.00001
Avg: $0.000006
#28
qwen/qwen3-235b-a22b-2507
qwen
58.75B active (235B total)
~401.8 mg (±25%)🟠
200.9 mg/602.6 mg
~772.6Wh (±25%)
386.3Wh/1158.9Wh
$0.0000/$0.0000
Avg: $0.0000
#29
qwen/qwen3-235b-a22b-thinking-2507
qwen
58.75B active (235B total)
~401.8 mg (±25%)🟠
200.9 mg/602.6 mg
~772.6Wh (±25%)
386.3Wh/1158.9Wh
$0.0000/$0.0000
Avg: $0.0000
#30
baidu/ernie-4.5-300b-a47b
baidu
75B active (300B total)
~482.6 mg (±25%)🟠
241.3 mg/723.9 mg
~1072.4Wh (±25%)
536.2Wh/1608.6Wh
$0.0000/$0.000001
Avg: $0.000001
#31
x-ai/grok-3-mini
x-ai
30B
Est.
~603.2 mg (±25%)🟠
301.6 mg/904.8 mg
~1340.5Wh (±25%)
670.3Wh/2010.8Wh
$0.0000/$0.000001
Avg: $0.0000
#32
anthropic/claude-opus-4
anthropic
400B
Est.
~618.4 mg (±25%)🟠
309.2 mg/927.6 mg
~2378.4Wh (±25%)
1189.2Wh/3567.6Wh
$0.000015/$0.000075
Avg: $0.000045
#33
x-ai/grok-3
x-ai
200B
Est.
~666.7 mg (±25%)🟠
333.3 mg/1.00 g
~1481.5Wh (±25%)
740.8Wh/2222.3Wh
$0.000003/$0.000015
Avg: $0.000009
#34
anthropic/claude-sonnet-4
anthropic
200B
Est.
~773 mg (±25%)🟠
386.5 mg/1.159 g
~2973.0Wh (±25%)
1486.5Wh/4459.5Wh
$0.000003/$0.000015
Avg: $0.000009
#35
openai/gpt-5
openai
400B
Est.
~1.637 g (±25%)🔴
818.6 mg/2.456 g
~5846.9Wh (±25%)
2923.4Wh/8770.3Wh
$0.000001/$0.00001
Avg: $0.000006
#36
mistralai/mistral-medium-3.1
mistralai
500B
Est.
~4.167 g (±25%)🔴
2.083 g/6.25 g
~9259.7Wh (±25%)
4629.8Wh/13889.5Wh
$0.0000/$0.000002
Avg: $0.000001
Environmental Impact Comparison

CO₂e Impact Categories

Excellent
< 0.001g CO₂e/1M tokens
Good
0.001-0.01g CO₂e/1M tokens
Fair
0.01-0.1g CO₂e/1M tokens
Concerning
0.1-1g CO₂e/1M tokens
Poor
> 1g CO₂e/1M tokens

Real-World Comparisons

1M tokens ≈
• ~750 pages of text
• ~1.5 hours of conversation
• ~500 code files
Carbon Footprint Context
• 0.001g CO₂e ≈ 5mm car drive
• 0.1g CO₂e ≈ 50cm car drive
• 1g CO₂e ≈ 5m car drive

Legend

Estimation Indicators

Est.
Estimated using hybrid system (regex + Hugging Face API + cost data)
Parameters: Extracted from model names, API metadata, or cost-based estimation
Architecture: Detected using pattern matching and API data
Rank: Calculated from environmental metrics

Are you an LLM provider?

±25% uncertainty range

Help us make your environmental data more accurate. We're constantly working to improve our measurements and would love to collaborate with you to ensure we're representing your models fairly and accurately.

Contact Our Research TeamLearn About Our Methodology

Important Disclaimer

These environmental impact estimates are based on our calibrated models using publicly available data and research. Actual emissions may vary significantly based on hardware configuration, data center efficiency, energy sources, and usage patterns. We encourage providers to share more accurate data to improve these estimates. This data is provided for informational purposes and should not be considered as definitive environmental impact assessments.