Leaderboard
CharXiv Reasoning
11 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.Muse Glimmer 30B158
- 2.Gemini 3.5 Flash Lite90.0
- 3.Gemini 3.5 Flash Lite87.6
- 4.Gemini 3.6 Flash59.6
- 5.Gemini 3.6 Flash56.8
| # | Model | Score | Best in /M | Best out /M | Pts / $ | Context |
|---|---|---|---|---|---|---|
| 🥇 | Kimi K3moonshotai | 91.3 | $2.50 | $13.50 | 17.4 | 1.05M |
| 🥈 | Gemini 3.6 Flashgoogle | 89.4 | $0.75 | $3.75 | 59.6 | 1.05M |
| 🥉 | Muse Spark 1.1meta | 88.4 | $1.25 | $4.25 | 44.2 | 1M |
| 4 | Fugu Ultrasakana | 86.6 | $5 | $30 | 7.7 | 1M |
| 5 | Gemini 3.6 Flashgoogle | 85.2 | $0.75 | $3.75 | 56.8 | 1.05M |
| 6 | Fugusakana | 85.1 | — | — | — | 1M |
| 7 | Gemini 3.5 Flashgoogle | 84.2 | $1.50 | $9 | 24.9 | 1.05M |
| 8 | Gemini 3.1 Pro Previewgoogle | 83.3 | $2 | $12 | 18.5 | 1.05M |
| 9 | Muse Glimmer 30Bmeta | 78.8 | $0.3 | $1.10 | 158 | 131K |
| 10 | Gemini 3.5 Flash Litegoogle | 76.5 | $0.3 | $2.50 | 90.0 | 1.05M |
| 11 | Gemini 3.5 Flash Litegoogle | 74.5 | $0.3 | $2.50 | 87.6 | 1.05M |
Cheapest scorer: Muse Glimmer 30B at $0.3 input /M.