Environmental ranking of AI assistants
Gemini has the highest environment score, 73 out of 100, ahead of Vibe (72) and Copilot (53). Frugality of the assistant in everyday use. 4 criteria scored from 0 to 3, scaled to 100 and weighted: a high score signals a frugal assistant.
Environment score of the 10 AI assistants, out of 100
Environment score out of 100, 4 weighted criteria; data as of 5 November 2026
- 1
GeminiGoogle · United States7383506783 - 2
VibeMistral AI · France7267757575 - 3
CopilotMicrosoft · United States5358335067 - 4
ClaudeAnthropic · United States4750673333 - 5
DeepSeekDeepSeek · China4558581733 - 6
PerplexityPerplexity · United States4250583317 - 7
ChatGPTOpenAI · United States4050333333 - 8
Meta AIMeta · United States3750171750 - 9
QwenAlibaba · China3333173350 - 10
GrokSpaceX Super Intelligence · United States1333000
Simple query, Advanced queries, Electricity, Approach: the detail is in the Table tab.
Sources for each score: Projet Celsius research of 4/11/2026, company publications, press, Stanford transparency index. CC BY 4.0 licence.
Criteria detail
Simple query
Consumption of a typical text query, on the default model served: figure published by the company, third-party estimate, failing that the size of the model.
- Gemini830.24 Wh and 0.26 mL of water measured by Google for the median text query of the Gemini app (May 2025), without naming the model.
- Vibe67Fast mode by default, reasoning on demand; 1.14 g CO₂e per half-page response (LCA, 2025).
- Copilot58Serves OpenAI models, unspecified; no figure for Copilot (the 0.31 Wh from Microsoft Research, in Joule in April 2026, apply to a frontier AI in general).
- DeepSeek58Very sparse models (13 billion active parameters for Flash) and shortened responses in October 2026; reasoning by default in the API.
- Claude50Default Sonnet: 0.83 Wh per typical conversation according to CLEER (Sonnet 5), against 0.50 Wh for Gemini 3.7 Flash; Haiku, 0.08 Wh, remains available.
- Perplexity50Each question triggers a web search and the reading of pages; model chosen without being shown (often Sonar, medium size).
- ChatGPT50Free: GPT-6 Luna, a small model; paid: Sol; all reason. Low-confidence CLEER estimates for their predecessors GPT-5.6: 0.17 and 1.36 Wh.
- Meta AI50Muse Spark announces 10 times less compute than Llama 4 Maverick (April 2026), with no energy figure; estimate: medium level.
- Qwen33Very large flagship models (Qwen3.8, 2,400 billion parameters, class F at compar:IA); the app's default model is not published.
- Grok33No figure published or estimated by a third party; an opaque Auto mode chooses the model that answers, and Grok 4.7, the frontier model, is offered as an option (Expert and Heavy modes).
Our view
Environment, one paragraph per assistantGemini. Of the 3 companies that publish a per-query figure, Google is the only one to publish the details of its measurement method: 0.24 Wh and 0.26 mL of water for the median text query of the Gemini app, measured in May 2025; Google does not say which model serves it. It announces 33 times less energy per query in one year and 65% carbon-free electricity hour by hour. Heavy uses are growing: video, Deep Research, 5 billion images in one month, and group emissions up 81% since 2019.
"Rankings" series on the Éclairages blog
All articles on AI →
Blog article ↗environmentWhich AIs pollute the most?10Grok139Qwen338Meta AI37environment score out of 100, the 3 lowest4 November 2026 · 12 min read
Blog article ↗ethicsWhich AI is the most ethical?1Claude772Copilot583ChatGPT52ethics score out of 1004 November 2026 · 13 min read
Blog article ↗Celsius scoreWhat are the best AIs of 2026?1Claude742Gemini713ChatGPT62Celsius score out of 1004 November 2026 · 11 min read
Blog article ↗useWhich AIs are the most used?1ChatGPT56.4 %2Gemini25.0 %3Claude9.6 %share of web traffic, September 20264 November 2026 · 12 min readContinue in the ranking
The full ranking →Comparison2 AI assistantsGemini or Vibe73 against 72environment score out of 100The top 2 of the category, on each score and each criterion.The comparison →
ClimateThe hub's climate topic: the summary, charts and figures.The topic →AssistantGemini71out of 100, 2nd out of 101st in environment: its scores, its plans, its models.The entry →The hub newsletter
This month's ranking, news on the impact of AI and one figure from the AI Impact Database. One newsletter a month at most, free of charge. Past issues
