Skip to content
LLM Pulse

Leaderboard

DeepSWE

12 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. 1.DeepSeek V4 Flash 0731518
  2. 2.GPT-5.6 Luna299
  3. 3.Gemini 3.7 Flash87.1
  4. 4.Gemini 3.6 Flash32.7
  5. 5.GPT-5.6 Terra30.9
#ModelScoreBest in /MBest out /MPts / $Context
🥇GPT-5.6 Solopenai
72.7
$2.50$1512.91.05M
🥈GPT-5.6 Terraopenai
69.6
$1$630.91.05M
🥉Claude Opus 5anthropic
68.8
$5$256.91M
4Kimi K3moonshotai
67.5
$2.50$13.5012.91.05M
5GPT-5.6 Lunaopenai
67.2
$0.1$0.62991.05M
6Gemini 3.7 Flashgoogle
65.3
$0.375$1.8887.11.05M
7Grok 4.5xai
62
$2$620.7500K
8DeepSeek V4 Flash 0731deepseek
54.4
$0.08$0.185181M
9Muse Spark 1.1meta
53.3
$1.25$4.2526.61M
10Grok 4.5xai
53
$2$617.7500K
11Gemini 3.6 Flashgoogle
49
$0.75$3.7532.71.05M
12GLM-5.2zhipuai
46.2
$1.40$4.4021.51M

Cheapest scorer: DeepSeek V4 Flash 0731 at $0.08 input /M.