Head to head
Qwen3 14BvsQwen2.5 7B Instruct
For tasks, about a month.
Verdict · measured basis
Qwen2.5 7B Instruct costs $0.0001 per solved task against
Qwen3 14B at $0.0012 — ▼ 19x cheaper.
Over 200 tasks that is $0.23 a month.
On output token price alone Qwen2.5 7B Instruct looks 1.2x cheaper. The ranking holds, but the margin changes.
Verified pricing, pass rate and cost per solved task for Qwen3 14B and Qwen2.5 7B Instruct| Figure | Qwen3 14B | Qwen2.5 7B Instruct |
|---|
| Input $/Mtok | $0.12 | $0.1 |
|---|
| Output $/Mtok | $0.24 | $0.2 |
|---|
| Cached input $/Mtok | $0.12 | $0.1 |
|---|
| Pass rate | 78% | 56% |
|---|
| Cost per solved task | $0.0012measured | $0.0001measured |
|---|
| Cost per month | $0.24 | $0.01 |
|---|
Cost per month against tasks per month · drag the marker
Solid measured · dashed modeled · dotted stale. The gap is the verdict.
Solvency Bench (solvency.dev) — first-party measurement · verified 2026-08-26 · prices from each provider's pricing page; measured rows ignore the tier · Methodology · All models at these settings