Models
The frontier, benchmarked
Every major AI model — context, price and arena score, head-to-head. The leaderboard of intelligence.
30 of 30
#ModelContext$/M inArena
1G
1000K$2.515012Gemini 3 Pro
google-deepmind
K
128K$0.1514543Kimi K2
G
1000K$1.2514444Gemini 2.5 Pro
google-deepmind
D
128K$0.5513895DeepSeek-R1
Q
262K$0.2513746Qwen3 235B
G
128K$0.0813657Gemma 3 27B
google-deepmind
C
200K$313328Claude Sonnet 4.5
anthropic
L
10000K$0.1813229Llama 4 Scout
meta-ai
M
256K$0.5610Mistral Large 3
mistral-ai
G
1048K$2—11GPT-4.1
openai
H
131K$0.05—12Hermes 4
nous-research
G
1000K$2.5—13GPT-5.4
openai
G
400K$0.75—14GPT-5.4 Mini
openai
G
1050K$5—15GPT-5.5
openai
G
256K$3—16Grok 4
xai
J
256K$2—17Jamba 1.6 Large
ai21-labs
L
128K——18Llama 3.3 70B
meta-ai
L
10000K——19Llama 4 Maverick
meta-ai
M
205K$0.26—20MiniMax M2
M
128K$0.4—21Mistral Medium 3
mistral-ai
C
200K$1—22Claude Haiku 4.5
anthropic
Q
262K$0.78—23Qwen3-Max
G
200K$0.35—24GLM-4.6
G
131K$3—25Grok 3
xai
M
128K$2—26Magistral
mistral-ai
P
16K$0.07—27Phi-4
C
200K$5—28Claude Opus 4.5
anthropic
C
256K$2.5—29Command A
cohere
D
128K$0.27—30DeepSeek-V3
G
1000K$0.5—Gemini 3 Flash
google-deepmind