Head to head
GLM 5.3 Flashvsgpt-oss-20b
For tasks, about a month.
Verdict · measured basis
gpt-oss-20b costs $0.0003 per solved task against
GLM 5.3 Flash at $0.0033 — ▼ 12x cheaper.
Over 200 tasks that is $0.61 a month.
On output token price alone gpt-oss-20b looks 1.9x cheaper. The ranking holds, but the margin changes.
Verified pricing, pass rate and cost per solved task for GLM 5.3 Flash and gpt-oss-20b| Figure | GLM 5.3 Flash | gpt-oss-20b |
|---|
| Input $/Mtok | $0.075 | $0.03 |
|---|
| Output $/Mtok | $0.25 | $0.13 |
|---|
| Cached input $/Mtok | $0.015 | $0.03 |
|---|
| Pass rate | 50% | 94% |
|---|
| Cost per solved task | $0.0033measured | $0.0003measured |
|---|
| Cost per month | $0.67 | $0.05 |
|---|
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