Leaderboard
SWE-Bench Multilingual
7 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 5 by pts per dollar
- 1.Laguna XS 2.1841
- 2.LongCat-2.0147
- 3.Nemotron 3 Ultra 550B A55B73.2
- 4.Grok 4.526.0
- 5.Qwen3.7 Max20.9
| # | Model | Score | Best in /M | Best out /M | Pts / $ | Context |
|---|---|---|---|---|---|---|
| 🥇 | Claude Opus 5anthropic | 89.5 | $5 | $25 | 8.9 | 1M |
| 🥈 | Qwen3.7 Maxalibaba | 78.3 | $2.50 | $7.50 | 20.9 | 1M |
| 🥉 | Claude Sonnet 5anthropic | 78.3 | $2 | $10 | 19.6 | 1M |
| 4 | Grok 4.5xai | 78 | $2 | $6 | 26.0 | 500K |
| 5 | LongCat-2.0meituan | 77.3 | $0.3 | $1.20 | 147 | 1M |
| 6 | Nemotron 3 Ultra 550B A55Bnvidia | 67.7 | $0.5 | $2.20 | 73.2 | 1M |
| 7 | Laguna XS 2.1poolside | 63.1 | $0.06 | $0.12 | 841 | 262K |
Cheapest scorer: Laguna XS 2.1 at $0.06 input /M.