Head to head
GPT-5.6 LunavsQwen2.5 72B Instruct
For tasks, about a month.
Verdict · measured basis
Qwen2.5 72B Instruct costs $0.0002 per solved task against
GPT-5.6 Luna at $2.65 — ▼ 16645x cheaper.
Over 200 tasks that is $530 a month.
On output token price alone Qwen2.5 72B Instruct looks 3.0x cheaper. The ranking holds, but the margin changes.
Verified pricing, pass rate and cost per solved task for GPT-5.6 Luna and Qwen2.5 72B Instruct| Figure | GPT-5.6 Luna | Qwen2.5 72B Instruct |
|---|
| Input $/Mtok | $0.2 | $0.36 |
|---|
| Output $/Mtok | $1.2 | $0.4 |
|---|
| Cached input $/Mtok | $0.02 | $0.36 |
|---|
| Pass rate | 57% | 53% |
|---|
| Cost per solved task | $2.65measured | $0.0002measured |
|---|
| Cost per month | $530 | $0.03 |
|---|
Cost per month against tasks per month · drag the marker
Solid measured · dashed modeled · dotted stale. The gap is the verdict.
Source: Artificial Analysis (artificialanalysis.ai) · verified 2026-08-21 · prices from each provider's pricing page; measured rows ignore the tier · Methodology · All models at these settings