Head to head
ME Llama 4 Maverickvs ME Llama 3.1 70B InstructFor light moderate heavy tasks, about a month.
Verdict · measured basis
Llama 4 Maverick costs $0.0002 per solved task against
Llama 3.1 70B Instruct at $0.0005 — ▼ 3.0x cheaper .
Over 200 tasks that is $0.07 a month.
On output token price alone Llama 3.1 70B Instruct looks 2.0x 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 Llama 4 Maverick and Llama 3.1 70B Instruct Figure Llama 4 Maverick Llama 3.1 70B Instruct Input $/Mtok $0.2 $0.4 Output $/Mtok $0.8 $0.4 Cached input $/Mtok $0.2 $0.4 Pass rate 69% 50% Cost per solved task $0.0002 measured $0.0005 measured Cost per month $0.04 $0.11
Cost per month against tasks per month · drag the marker
Monthly cost against tasks per month Log-log lines for Llama 4 Maverick (measured, $0.0002 per solved task), Llama 3.1 70B Instruct (measured, $0.0005 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 Llama 4 Maverick · MEASURED · $0.0002 per solved task $0.04 Llama 4 Maverick MEASURED Llama 3.1 70B Instruct · MEASURED · $0.0005 per solved task $0.11 Llama 3.1 70B Instru… MEASURED Monthly cost against tasks per month Log-log lines for Llama 4 Maverick (measured, $0.0002 per solved task), Llama 3.1 70B Instruct (measured, $0.0005 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 Llama 4 Maverick · MEASURED · $0.0002 per solved task $0.04 Llama 3.1 70B Instruct · MEASURED · $0.0005 per solved task $0.11 — Llama 4 Maverick · MEASURED — Llama 3.1 70B Instruct · 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