Directories / Models / Cheapest Reason

Cheapest models for reason

Ranked by input $/Mtoken, then output. Hard reasoning and planning workloads.

1
Claude Sonnet 5 (Free)
Anthropic · 1,000,000 context

Input

$0/M

Output

$0/M
2
GLM 5.2 (Free)
Z Ai · 1,000,000 context

Input

$0/M

Output

$0/M
3
Kimi K2.7 Code (Free)
Moonshotai · 262,144 context

Input

$0/M

Output

$0/M
4
glm-4.7
Novita · 205,000 context

Input

$0/M

Output

$0/M
5
kimi-k2-thinking
Novita · 256,000 context

Input

$0/M

Output

$0/M
6
minimax-m2.1
Novita · 205,000 context

Input

$0/M

Output

$0/M
7
Step 3.7 Flash (Free)
Stepfun · 256,000 context

Input

$0/M

Output

$0/M
8
GLM-4.6
Zhipuai · 202,752 context

Input

$0/M

Output

$0/M
9
gpt-oss-120b (free)
OpenAI · 131,072 context

Input

$0/M

Output

$0/M
10
glm-4.7-flash
Novita · 200,000 context

Input

$0/M

Output

$0/M
11
GLM 4.7 Flash (Free)
Z Ai · 200,000 context

Input

$0/M

Output

$0/M
12
Cohere Command A
Cohere · 128,000 context

Input

$0/M

Output

$0/M
13
GLM-4.5
Zhipuai · 131,072 context

Input

$0/M

Output

$0/M
14
Grok 3
xAI · 128,000 context

Input

$0/M

Output

$0/M
15
glm-4.6v
Novita · 131,000 context

Input

$0/M

Output

$0/M
16
Mistral Small 3.1
Mistral Ai · 128,000 context

Input

$0/M

Output

$0/M
17
Mistral Medium 3 (25.05)
Mistral Ai · 128,000 context

Input

$0/M

Output

$0/M
18
nvidia-nemotron-nano-9b-v2
NVIDIA · 131,072 context

Input

$0/M

Output

$0/M
19
Phi-4-mini-instruct
Microsoft · 128,000 context

Input

$0/M

Output

$0/M
20
Phi-4-Reasoning
Microsoft · 128,000 context

Input

$0/M

Output

$0/M
21
Phi-4-multimodal-instruct
Microsoft · 128,000 context

Input

$0/M

Output

$0/M
22
Mistral Large 24.11
Mistral Ai · 128,000 context

Input

$0/M

Output

$0/M
23
Cohere Command R
Cohere · 128,000 context

Input

$0/M

Output

$0/M
24
LFM2.5-1.2B-Thinking (free)
Liquid · 32,768 context

Input

$0/M

Output

$0/M
25
AI21 Jamba 1.5 Large
Ai21 Labs · 256,000 context

Input

$0/M

Output

$0/M
26
AI21 Jamba 1.5 Mini
Ai21 Labs · 256,000 context

Input

$0/M

Output

$0/M
27
Baidu: CoBuddy (free)
Baidu · 131,072 context

Input

$0/M

Output

$0/M
28
qwen3-235b-2507-cs
Cerebras · 1 context

Input

$0/M

Output

$0/M
29
qwen3-32b-cs
Cerebras · 1 context

Input

$0/M

Output

$0/M
30
JAIS 30b Chat
Core42 · 8,192 context

Input

$0/M

Output

$0/M
31
Deepseek/Deepseek-Math-V2
DeepSeek · 160,000 context

Input

$0/M

Output

$0/M
32
Deepseek/DeepSeek-V3.2
DeepSeek · 128,000 context

Input

$0/M

Output

$0/M
33
DeepSeek/DeepSeek-V3.1-Terminus-Thinking
DeepSeek · 128,000 context

Input

$0/M

Output

$0/M
34
DeepSeek/DeepSeek-V3.2-Exp-Thinking
DeepSeek · 128,000 context

Input

$0/M

Output

$0/M
35
Gemma 4 31B IT FP8
Google · 262,144 context

Input

$0/M

Output

$0/M
36
Kilo Auto Free
Kilo Auto · 204,800 context

Input

$0/M

Output

$0/M
37
LongCat Flash Thinking 2601
Meituan · 32,768 context

Input

$0/M

Output

$0/M
38
Llama-3.2-11B-Vision-Instruct
Meta · 128,000 context

Input

$0/M

Output

$0/M
39
Llama-3.2-90B-Vision-Instruct
Meta · 128,000 context

Input

$0/M

Output

$0/M
40
Llama-3.3-70B-Instruct
Meta · 128,000 context

Input

$0/M

Output

$0/M

Full reasoning leaderboard → All models