Head to head
QW vs For light moderate heavy
Verdict · measured basis
Qwen3 VL 32B Instruct costs $0.0001 per solved task against
gpt-oss-20b at $0.0003 — ▼ 2.7x cheaper .
Over 200 tasks that is $0.03 a month.
On output token price alone gpt-oss-20b looks 3.2x cheaper. The ranking reverses once you divide by pass rate — the cheaper tokens do not win the task.
Copy link Save scenario Export compare table Planned
Verified pricing, pass rate and cost per solved task for Qwen3 VL 32B Instruct and gpt-oss-20b Figure Qwen3 VL 32B Instruct gpt-oss-20b Input $/Mtok $0.104 $0.03 Output $/Mtok $0.416 $0.13 Cached input $/Mtok $0.104 $0.03 Pass rate 67% 94% Cost per solved task $0.0001 measured $0.0003 measured Cost per month $0.02 $0.05
Cost per month against tasks per month · drag the marker
Monthly cost against tasks per month Log-log lines for Qwen3 VL 32B Instruct (measured, $0.0001 per solved task), gpt-oss-20b (measured, $0.0003 per solved task) from 10 to 100,000 tasks a month. Marker at 200 tasks. 10 100 1k 10k 100k $0.001 $0.01 $0.1 $1 $10 TASKS / MONTH (log) → 200 tasks Qwen3 VL 32B Instruct · MEASURED · $0.0001 per solved task $0.02 Qwen3 VL 32B Instruct MEASURED gpt-oss-20b · MEASURED · $0.0003 per solved task $0.05 gpt-oss-20b MEASURED Monthly cost against tasks per month Log-log lines for Qwen3 VL 32B Instruct (measured, $0.0001 per solved task), gpt-oss-20b (measured, $0.0003 per solved task) from 10 to 100,000 tasks a month. Marker at 200 tasks. 10 100 1k 10k 100k $0.001 $0.01 $0.1 $1 $10 TASKS / MONTH (log) → 200 tasks Qwen3 VL 32B Instruct · MEASURED · $0.0001 per solved task $0.02 gpt-oss-20b · MEASURED · $0.0003 per solved task $0.05 — Qwen3 VL 32B Instruct · MEASURED — gpt-oss-20b · MEASURED 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