AI 
App / editor
ChatGPT (agents, GPTs, Codex)
3939/100 · Poor100🌿 Env.▾
Overall score · 39/100
entry paid plan with the feature (Plus 20 $/mois) ≈ $20; agents median $20 (ratio 1.00) — official prices checked 10/16/2026 · Contact us
🌿 Environmental score · 100/100
estimate ≤ 0.24 Wh / text prompt (proxy) (impact-estimates) · Contact us
estimates, not measurements
How it's computed : Overall score: weighted mean of 1–5 sub-scores scaled to 100 — Quality 45, Price 30, Security 25. Separate environmental score: Electricity 50, Water 50 (estimates, CO₂ derived from electricity). Missing sub-score = excluded and weights renormalized ('partial data'). Never influenced by affiliation. Methodology
The short answer
Is ChatGPT (agents, GPTs, Codex) for you?
An assistant that chats, searches, writes code and can do tasks for you.
Yes, if you want:
Best AI coding agent for developers
Codex, OpenAI's coding assistant, is included in ChatGPT (more usage from Plus). Its model is 21st in the public ranking, hence an average score. (#5 in our picks)
Best agent builder
Custom assistants, scheduled tasks and agent mode from Plus ($20/mo). Its best model is only 21st in the public ranking, hence a lower quality score. (#5 in our picks)
Best for vibe coding
The Codex coding assistant is included (limited on Free, more from Plus). Lower quality score: its model is only 21st in the public ranking. (#7 in our picks)
To get started
How much?
$20 /mo
Plus plan.
Or free to try.
Official price checked 2026-10-16, excl. tax.
Overall score (quality, price, security)
Poor
39 out of 100
🌿 Environmental score (electricity, water)
Excellent
100 out of 100 · estimate
Our take
In one sentence
Average quality, affordable price.
Our score never depends on affiliate links.
Other options
If it's not right for you
- Cheaper
Gemini (Gems, Canvas)
Two assistants that code for you (Jules and Antigravity) in Google AI Pro, with the first-ranked model. Less built into a code editor than Cursor.
From $4.99/mo · free plan
- Cheaper
Cursor
The go-to code editor with a built-in assistant: it completes code, edits several files and understands your project. Plan limits change often.
From $8/mo · free plan
- Cheaper
GitHub Copilot
The easiest if your team already works on GitHub and in VS Code. Less bold on big changes done on its own.
From $10/mo · free plan
Ready to try?
There is a free plan to try it.
Go to ChatGPT (agents, GPTs, Codex)Overall
39/100
Environment
100/100
estimate, separate
Entry price
$20/mo
Plus
Sub-scores and weights
| Criterion | Weight | /100 | Basis |
|---|---|---|---|
| Quality | 45 % | 25 | LMArena Text #21 (gpt-5.6-sol-xhigh), checked 10/16/2026 |
| Price | 30 % | 50 | entry paid plan with the feature (Plus 20 $/mois) ≈ $20; agents median $20 (ratio 1.00) — official prices checked 10/16/2026 |
| Security | 25 % | 50 | C (ChatGPT Free/Plus, securite-donnees.csv) |
| Overall score | 100 % | 39 |
🌿 Environmental score (separate, not part of the overall score)
| Criterion | Weight | /100 | Estimate basis |
|---|---|---|---|
| Electricity | 50 % | 100 | estimate ≤ 0.24 Wh / text prompt (proxy) (impact-estimates) |
| Water | 50 % | 100 | estimate ≤ 0.432 mL / text prompt (proxy) (impact-estimates) |
| Environmental score | 100 % | 100 | Estimates, not measurements. |
Methodology · never influenced by affiliation
Pricing, plan by plan
| Plan | Official price | Credits / mo | ≈ voice minutes | Commercial |
|---|---|---|---|---|
| Free | — | — | — | No |
| Go | see website | — | — | — |
| Plus | $20/mo (custom GPTs, scheduled tasks, Codex) | — | — | — |
| Pro | 3 tiers, including $200/mo | — | — | — |
| Business / Enterprise | per user, see site | — | — | — |
Source: help.openai.com · checked 2026-10-16 · USD excl. tax. Minutes estimated at ≈1,000 credits per voice minute.
Rankings
Official sources
| Topic | URL | Date |
|---|---|---|
| Pricing | help.openai.com/en/articles/6950777-what-is-chatgpt-plus | 2026-10-16 |
Date = access date (pricing, ToS, aggregators) or publication date (versions).
Estimated impact
Footprint (estimate)
Estimate — low/high range · not a service measurement
- Electricity: 0.24 Wh / text prompt (proxy)
- Water: 0.43 mL / text prompt (proxy)
- CO₂: 0.105 g / text prompt (proxy)
Where servers run: OpenAI · region not publicly disclosed (region not disclosed, default average used) · hosting source · grid world average (default) (436 gCO₂/kWh) · grid source
All electricity, water and CO₂ figures are estimates (never direct service measurements), with a low/high range. Verified sources are listed on /methodologie. Method