Head to head
Qwen3 8BvsLlama 3.1 8B Instruct
For tasks, about a month.
Verdict · measured basis
Llama 3.1 8B Instruct costs $0.0000 per solved task against
Qwen3 8B at $0.0043 — ▼ 113x cheaper.
Over 200 tasks that is $0.86 a month.
On output token price alone Llama 3.1 8B Instruct looks 5.7x cheaper. The ranking holds, but the margin changes.
Verified pricing, pass rate and cost per solved task for Qwen3 8B and Llama 3.1 8B Instruct| Figure | Qwen3 8B | Llama 3.1 8B Instruct |
|---|
| Input $/Mtok | $0.117 | $0.05 |
|---|
| Output $/Mtok | $0.455 | $0.08 |
|---|
| Cached input $/Mtok | $0.117 | $0.025 |
|---|
| Pass rate | 57% | 42% |
|---|
| Cost per solved task | $0.0043measured | $0.0000measured |
|---|
| Cost per month | $0.87 | $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