NOUMENOpen app
NOUMEN

OpenAI · Balanced

OpenAI o3

Added Apr 16, 2025

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

OpenAI o3 is a reasoning model for complex analytical, coding, mathematical, and scientific tasks.

  • Reasoning
Input
Text
Output
Text
Reasoning effort
low · medium · high
Discovery tier
Balanced

Pricing

USD per 1M tokens
Input
$2.00
Output
$8.00
Cache read
$1.00

Reference prices as published by the model provider. API example

Specifications

Context window
200K
Max output
100K
Average output (7 days)
205 tokens
Longer than 22% of 468 models

Side by side

OpenAI o3 compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

OpenAI o3OpenAIThis modelQwen3.8 27B TEEQwenRemove
ProviderOpenAIQwen
Discovery tierBalancedOpen weights
AddedApr 16, 2025Aug 24, 2026
Limits
Context window200K262.14KBest
Max output100K262.14KBest
Average output205 tokens301 tokens
Pricing, USD per 1M tokens
Input$2.00$0.40Best
Output$8.00$3.00Best
Cache read$1.00$0.15Best
Performance
Request success (24h)—100.0%
Benchmark indexes
Intelligence Index20.220.2
Coding Index—44.6
Math Index88.3—
Capabilities
ReasoningYesYes
ToolsNoYes
VisionNoYes
Structured outputNoYes
Video inputNoYes
Reasoning effortlow, medium, high—

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • OpenAI o3
  • Qwen3.8 27B TEE

Intelligence Index

  • OpenAI o320.2
  • Qwen3.8 27B TEE20.2

OpenAI o3: Better than 50% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 44.6

  • Math Index

    Composite of the math evaluations

    • OpenAI o3: 88.3
    • Qwen3.8 27B TEE: —

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 12.4%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 1.5%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 898 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 1,074 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 5.2%

  • AA-LCRv1.1

    Reasoning across long documents

    • OpenAI o3: 74.7%Top
    • Qwen3.8 27B TEE: 69.3%

  • MLCR-AA

    Long-context reasoning over medical records

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 10.0%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • OpenAI o3: 20.1%Top
    • Qwen3.8 27B TEE: 12.1%

  • GPQA Diamond

    Graduate-level science questions

    • OpenAI o3: 82.7%Top
    • Qwen3.8 27B TEE: 81.8%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • OpenAI o3: 85.3%
    • Qwen3.8 27B TEE: —

  • IFBench

    Following precise output instructions

    • OpenAI o3: 71.4%
    • Qwen3.8 27B TEE: —

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 0.0%

  • SciCode

    Scientific computing in Python

    • OpenAI o3: —
    • Qwen3.8 27B TEE: 36.2%

  • LiveCodeBench

    Recent competitive programming problems

    • OpenAI o3: 80.8%
    • Qwen3.8 27B TEE: —

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • OpenAI o3: 88.3%
    • Qwen3.8 27B TEE: —

  • AIME

    Competition mathematics

    • OpenAI o3: 90.3%
    • Qwen3.8 27B TEE: —

  • MATH-500

    Competition math problems

    • OpenAI o3: 99.2%
    • Qwen3.8 27B TEE: —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • OpenAI o3: 37.1%
    • Qwen3.8 27B TEE: —

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • OpenAI o3: 80.7%
    • Qwen3.8 27B TEE: —

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • OpenAI o3: 74.7%Top
    • Qwen3.8 27B TEE: 69.3%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/o3",
    "reasoning_effort": "medium",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="openai/o3",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

OpenAI

Other OpenAI models, starting with the same family.