Leaderboard
GPQA Diamond
17 models scored · metric: accuracy. 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.GPT-5.6 Luna410
- 2.GPT-5.4 nano202
- 3.LongCat-2.0169
- 4.Muse Glimmer 30B167
- 5.MiMo-V2.5-Pro159
| # | Model | Score | Best in /M | Best out /M | Pts / $ | Context |
|---|---|---|---|---|---|---|
| 🥇 | Fugusakana | 95.5 | — | — | — | 1M |
| 🥈 | Fugu Ultrasakana | 95.5 | $5 | $30 | 8.5 | 1M |
| 🥉 | GPT-5.6 Solopenai | 94.6 | $2.50 | $15 | 16.8 | 1.05M |
| 4 | GPT-5.4 Proopenai | 94.4 | $27 | $160 | 1.6 | 1.05M |
| 5 | Gemini 3.1 Pro Previewgoogle | 94.3 | $2 | $12 | 21.0 | 1.05M |
| 6 | Claude Opus 4.7anthropic | 94.2 | $5 | $25 | 9.4 | 1M |
| 7 | GPT-5.5openai | 93.6 | $4.55 | $27.27 | 9.2 | 1.05M |
| 8 | GPT-5.6 Terraopenai | 92.9 | $1 | $6 | 41.3 | 1.05M |
| 9 | GPT-5.4openai | 92.8 | $2.20 | $14 | 18.0 | 1.05M |
| 10 | Qwen3.7 Maxalibaba | 92.4 | $2.50 | $7.50 | 24.6 | 1M |
| 11 | GPT-5.6 Lunaopenai | 92.3 | $0.1 | $0.6 | 410 | 1.05M |
| 12 | GLM-5.2zhipuai | 91.2 | $1.40 | $4.40 | 42.4 | 1M |
| 13 | LongCat-2.0meituan | 88.9 | $0.3 | $1.20 | 169 | 1M |
| 14 | GPT-5.4 miniopenai | 88 | $0.68 | $4 | 58.3 | 400K |
| 15 | MiMo-V2.5-Proxiaomi | 86.6 | $0.435 | $0.87 | 159 | 1.05M |
| 16 | Muse Glimmer 30Bmeta | 83.5 | $0.3 | $1.10 | 167 | 131K |
| 17 | GPT-5.4 nanoopenai | 82.8 | $0.18 | $1.10 | 202 | 400K |
Cheapest scorer: GPT-5.6 Luna at $0.1 input /M.