Skip to content
LLM Pulse

Leaderboard

SWE-Atlas Test Writing

12 models scored · metric: score. 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.Kimi K2.536.8
  2. 2.MiniMax-M2.535.4
  3. 3.Gemini 3 Flash Preview26.9
  4. 4.GLM-518.5
  5. 5.GPT-5.3 Codex8.8
#ModelScoreBest in /MBest out /MPts / $Context
🥇GPT-5.4openai
44.36
$2.20$148.61.05M
🥈GPT-5.5openai
42.59
$4.55$27.274.21.05M
🥉GPT-5.4openai
40
$2.20$147.81.05M
4GPT-5.3 Codexopenai
38.98
$1.60$138.8400K
5Claude Opus 4.6anthropic
36.67
$5$253.71M
6Claude Opus 4.6anthropic
36.08
$5$253.61M
7Claude Sonnet 4.6anthropic
31.76
$3$155.31M
8Gemini 3 Flash Previewgoogle
30.3
$0.5$326.91.05M
9Gemini 3.1 Pro Previewgoogle
29.84
$2$126.61.05M
10GLM-5zhipuai
28.74
$1$3.2018.5205K
11Kimi K2.5moonshotai
25.77
$0.3$1.9036.8262K
12MiniMax-M2.5minimax
18.6
$0.3$1.2035.4205K

Cheapest scorer: Kimi K2.5 at $0.3 input /M.