Qwen3.5 (397B-A17B) Environmental Impact
Flagship open-source MoE model with hybrid thinking — Apache 2.0
Energy per query
1.1 Wh
4x more than a Google search (0.3 Wh)
CO2 per query
0.58 g
China grid (525 gCO₂/kWh)
Water per query
4 mL
~244 queries to fill 1 litre
Processing location
Alibaba Cloud (China) / self-hosted
Provider
Alibaba
Category
Text / Chat
Grid carbon intensity
525 g CO2/kWh (33% renewable)
How does Qwen3.5 (397B-A17B) compare?
Detailed Breakdown
Energy Consumption
Qwen3.5 is one of the most capable open-source models available — a 397B-parameter Mixture-of-Experts design that activates only ~17B parameters per token, with native vision, 262K context, and hybrid thinking/non-thinking modes. At ~1.1 Wh per non-thinking query, the sparse activation keeps it competitive with much smaller dense models. Apache 2.0 licensed.
Power Source & Carbon
Available via Alibaba Cloud or self-hosted. Alibaba Cloud operates data centres in China (525 g CO2/kWh) and internationally. Being open-source, it's widely deployed on lower-carbon infrastructure globally.
Water Usage
At ~4.1 mL per query on Chinese infrastructure. Self-hosted globally, water varies by location.
About Qwen3.5 (397B-A17B)
Qwen3.5 (397B-A17B) is an open-source text and chat model from Alibaba, released in February 16, 2026, that runs well below the category average for energy consumption at 1.1 Wh per query. Because its weights are publicly available, it can be self-hosted on any infrastructure — meaning its carbon footprint depends entirely on where and how you choose to run it. At 397B parameters, it flagship open-source moe model with hybrid thinking — apache 2.0.
These figures are estimates derived from hardware specifications and API benchmarks — Alibaba has not published official energy data for Qwen3.5 (397B-A17B). Actual consumption may vary significantly depending on batching, quantisation, and infrastructure optimisations that we cannot observe from outside.
Qwen3.5 (397B-A17B) in Context
Your yearly Qwen3.5 (397B-A17B) footprint
At 25 queries per day, your annual Qwen3.5 (397B-A17B) usage consumes 10.0 kWh — roughly what a fridge uses in a month. That produces 5.3 kg of CO₂.
Key Insights
What does your Qwen3.5 (397B-A17B) usage cost the planet?
Use our calculator to estimate your personal environmental footprint based on how often you use Qwen3.5 (397B-A17B).
Calculate My ComputeFrequently Asked Questions
How much energy does Qwen3.5 (397B-A17B) use per query?
Each Qwen3.5 (397B-A17B) query consumes approximately 1.1 Wh of energy. This is 4x more than a traditional Google search (~0.3 Wh).
What is Qwen3.5 (397B-A17B)'s carbon footprint?
Based on the carbon intensity of Alibaba Cloud (China) / self-hosted, each query produces approximately 0.58 g of CO2. The grid in this region has a carbon intensity of 525 g CO2/kWh with 33% renewable energy.
How much water does Qwen3.5 (397B-A17B) use?
Each query consumes approximately 4 mL of water, primarily used for cooling the data centers that process the request.
How does Qwen3.5 (397B-A17B) compare to a Google search?
A Qwen3.5 (397B-A17B) query uses 4x more than a Google search in terms of energy. A Google search uses approximately 0.3 Wh, while Qwen3.5 (397B-A17B) uses 1.1 Wh.
Technical Details
Architecture
Mixture-of-Experts Transformer (397B total, ~17B active) with hybrid thinking
Parameters
397B
Context window
262,000 tokens
Release date
2026-02-16
Open source
Yes
Training data cutoff
2026-01