Skip to content
LLM Pulse

Leaderboard

Terminal-Bench Hard

22 models scored · metric: success 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.Step 3.5 Flash 2603217
  2. 2.Step 3.5 Flash192
  3. 3.Step 3.7 Flash85.5
  4. 4.Mistral Small 466.3
  5. 5.GLM-4.5-Air48.2
#ModelScoreBest in /MBest out /MPts / $Context
🥇GPT-5-Codexopenai
37.9
$1.10$912.3400K
🥈Step 3.7 Flashstepfun
35.6
$0.185$1.1185.5256K
🥉Step 3.5 Flash 2603stepfun
32.6
$0.1$0.3217256K
4Step 3.5 Flashstepfun
27.3
$0.09$0.3192256K
5Gemini 2.5 Progoogle
26.5
$0.87$711.01.05M
6GLM-4.6zhipuai
25
$0.6$2.2025.0205K
7GLM-4.5zhipuai
22
$0.6$2.2022.0131K
8Qwen3 Maxalibaba
20.5
$1.20$68.5262K
9GLM-4.5-Airzhipuai
20.5
$0.2$1.1048.2131K
10Devstral 2mistral
18.9
$0.4$223.6262K
11Mistral Small 4mistral
17.4
$0.15$0.666.3256K
12Mistral Large 3mistral
15.9
$0.5$1.5021.2262K
13Qwen3-Coder 30B-A3B Instructalibaba
15.2
$0.45$2.2516.9262K
14Gemini 2.5 Flashgoogle
13.6
$0.21$1.8022.41.05M
15GPT-4o (2024-08-06)openai
8.3
$2.50$101.9128K
16GPT-4o (2024-11-20)openai
8.3
$2.50$101.9128K
17DeepSeek-R1deepseek
6.1
$0.7$2.505.3128K
18Mistral Large 2.1mistral
6.1
$2$62.0131K
19GLM-4.5Vzhipuai
5.3
$0.6$1.805.964K
20Gemini 2.5 Flash-Litegoogle
4.5
$0.1$0.145.01.05M
21Mistral Medium 3mistral
3.8
$0.4$24.7131K
22Llama-3.3-70B-Instructmeta
3
$0.22$0.510.3128K

Cheapest scorer: Step 3.5 Flash at $0.09 input /M.