OpenAIo1Arena #79Dec 17, 2024

o1

The latest and strongest model family from OpenAI, o1 is designed to spend more time thinking before responding. The o1 model series is trained with large-scale reinforcement learning to reason using chain of thought. The o1 models are optimized for math, science, programming, and other STEM-related tasks. They consistently exhibit PhD-level accuracy on benchmarks in physics, chemistry, and biology. Learn more in the launch announcement.

Context Window
200K
tokens
Max Output
100K
tokens
Released
Dec 17, 2024
Arena Rank
#79
of 305 models
Output Speed
52
tokens/sec
Time to First Token
22s
TTFT

Capabilities

👁Vision
🧠Reasoning
🔧Tool Calling
Prompt Caching
🖥Computer Use
🎨Image Generation

Supported Parameters

Max Tokens
Output length limit
Response Format
JSON mode / structured output
Seed
Deterministic outputs
structured_outputs
Tool Choice
Control tool usage
Tool Calling
Function calling support

Pricing Comparison

RouterInput / 1MOutput / 1MCached Input / 1M
Requesty$15.00$60.00$7.50
OpenRouter$15.00$60.00$7.50
Vercel AI$15.00$60.00
Martian$15.00$60.00$7.50

Benchmarks

Artificial Analysis
Intelligence IndexArtificial Analysis
68/100
Coding IndexArtificial Analysis
60/100
Math IndexArtificial Analysis
93/100
MMLU-PROArtificial Analysis
0.81/1
GPQA DiamondArtificial Analysis
0.78/1
MATH-500Artificial Analysis
0.96/1
AIME 2024Artificial Analysis
0.83/1
Humanity's Last ExamArtificial Analysis
0.095/1
LiveCodeBenchArtificial Analysis
0.63/1
Aider Code Editing
Aider EditingAider Code Editing
57.9/100
Aider EditingAider Code Editing
65.4/100

Model IDs

Requestyopenai/o1:high
OpenRouteropenai/o1

Tags

visionreasoningtool-callingcaching
Compare with another model

Related Models