One job, one card, one API
Qwen: Qwen3 14B on GeForce RTX 4080 SUPER
The sums below use the “Chatbot for a small team” example: 4,200 jobs a month, 1,500 tokens in, 400 out. 8 people, 25 questions on each of 21 working days. 1,500 tokens in (the question plus a short history) and 400 tokens out. A worked example, not a log from a real team.
Not enough yet
The API comes to £0.74 a month for this job. There is no speed for running it locally, so the electricity and the break-even are left blank.
Memory. Qwen: Qwen3 14B against GeForce RTX 4080 SUPER's 16 GB, estimated from the Hub's published weight size across the common quants: BF16 29.5 GB no · Q8_0 15.7 GB yes · Q6_K 12.2 GB yes · Q5_K_M 10.5 GB yes · Q4_K_M 9 GB yes · IQ4_XS 7.8 GB yes · Q3_K_M 7.2 GB yes · Q2_K 5.5 GB yes. At Q8_0, about 15.7 GB, 0.3 GB of the card's 16 GB is left for the context cache. Rule of thumb, not a measured file size.
- API this month
- £0.74
- Electricity this month
- Blank
- GPU purchase
- Not on file
- Break-even
- Needs a speed
- API cost is input tokens and output tokens, priced per million, times jobs per month. No caching, batch discount, or free tier.
- Pound figures use the stored exchange rate. They are not a quote from the company that trained the model.
- No tokens-per-second figure, so the electricity line is blank. Type one in if you have measured it.
- API price: £0.0745 in and £0.164 out per million tokens, from $0.100 and $0.220 at 0.74501 dollars to the pound on 2026-11-10. OpenRouter feed 2026-11-11.
- GeForce RTX 4080 SUPER: 16 GB GDDR6X. Total graphics power: 320 W. Specs read 2026-11-11.
- No sourced speed for this pair, and none was typed in.
- Upper estimate, based on the official total graphics power of 320 W.
OpenRouter prices · Exchange rate · GPU spec · Ofgem unit rate