Leaderboard
SWE Marathon
3 models scored · metric: resolve rate. Prices are the cheapest listed across all serving providers. Pts/$ = score ÷ blended price (3×input + output) / 4 — how much score each dollar buys.
Best value · top 3 by pts per dollar
| # | Model | Score | Best in /M | Best out /M | Pts / $ | Context |
|---|---|---|---|---|---|---|
| 🥇 | Kimi K3moonshotai | 42 | $2.50 | $13.50 | 8.0 | 1.05M |
| 🥈 | Grok 4.5xai | 29 | $2 | $6 | 9.7 | 500K |
| 🥉 | GLM-5.2zhipuai | 13 | $1.40 | $4.40 | 6.0 | 1M |
Cheapest scorer: GLM-5.2 at $1.40 input /M.