Best for Chat & Knowledge

All models2292Chat & Knowledge558Coding381Agents & Tool Use541Writing & Language360Vision & Documents395Image Generation & Editing183Video Generation148Speech & Voice358Embeddings & Retrieval671Decision models118Music & Sound29
ModelMathArena · ArXivMath January 2026Per taskPriceSpeedSupportsContext
GPT-5.4OpenAI
76xhigh
—$2.50 / $151271.05M
DeepSeek V4 Pro 0813DeepSeekOpen
74max
—$1.32 / $3.96—1M
GPT-5.5OpenAI
74xhigh
$2.63$5 / $301131.05M
Claude Opus 4.6Anthropic
73high
—$5 / $25451M
Kimi K2.6Moonshot AIOpen
72unstated
—$0.95 / $4—256K
Gemini 3.1 Pro PreviewGoogle
71unstated
—$2 / $12—1M
GPT-5.2OpenAI
68unstated
—$1.75 / $14—400K
GLM 5.1Z.aiOpen
65unstated
—$1.40 / $4.40—203K
Kimi K2.5Moonshot AIOpen
63unstated
—$0.60 / $3—256K
Gemini 3 Flash PreviewGoogle
58unstated
—$0.50 / $3—1M
Qwen3.5 397B A17BAlibaba (Qwen)Open
55unstated
—$0.60 / $3.60—262K
GLM 5Z.aiOpen
54unstated
—$1 / $3.20—203K
Qwen 3.5 27BAlibaba (Qwen)Open
53unstated
—$0.30 / $2.40—262K
Claude Opus 4.7Anthropic
52xhigh
—$5 / $25—1M
Qwen 3.5 35B A3BAlibaba (Qwen)Open
50unstated
—$0.25 / $2—262K
Qwen 3.5 9BAlibaba (Qwen)Open
45unstated
—$0.08 / $0.13—262K
Qwen3.5 4BAlibaba (Qwen)Open
38unstated
—$0.03 / $0.15—262K
lQED Nanolm-proversOpen
37
—$0.40 / $0.80—262K