LLM·Dex
Newsletter
ProprietaryOpenAIflagshiptextvision

o4

OpenAI's late-2025 standalone reasoning model, an evolution of o3 with deeper chain-of-thought and stronger multimodal reasoning.

Updated

Quick facts

Released
Dec 2025
Context
200K tokens
Output / 1M
Pricing not published
License
Proprietary

About o4

o4 continued OpenAI's reasoning-model line after GPT-5 unified reasoning into the chat model. It exists for use cases where dedicated, slower thinking is preferred over the routed-reasoning approach of GPT-5, typically high-stakes math, science, and planning workloads.

Like o3, o4's value is on hard problems, not chat. Routine queries are slower and more expensive than necessary; complex multi-step reasoning is meaningfully better than what GPT-5 produces by default.

Benchmarks

Published scores from OpenAI's model card or independent leaderboards. We do not publish numbers we cannot source, see methodology.

HumanEval
,
Python coding pass@1
MMLU
,
Broad academic knowledge
GPQA
,
Graduate-level reasoning
SWE-bench
,
Real software-engineering tasks
Benchmark scores not yet available. We only publish numbers we can source from official model cards or independent leaderboards, see methodology.

Capabilities

Strengths

  • Exceptional performance on hard math and reasoning benchmarks
  • Good at multi-step planning and verification
  • Strong scientific reasoning

Tracked weaknesses

  • Slow, reasoning tokens take real wall-clock time
  • Expensive on reasoning-heavy queries
  • Overkill for chat or simple tasks

Pricing

Per-million-token rates as published by OpenAI.

Per-token pricing not yet published for o4. Check the official provider site for current tiers.

Call o4 from your code

Drop-in snippet for the OpenAI SDK. Set your API key in the environment and run.

typescript
import OpenAI from "openai";

const client = new OpenAI({
  // Use OPENAI_API_KEY for OpenAI, or your provider's key + baseURL.
  apiKey: process.env.OPENAI_API_KEY,
});

const completion = await client.chat.completions.create({
  model: "o4",
  messages: [
    { role: "user", content: "What's the time complexity of quicksort?" },
  ],
});

console.log(completion.choices[0].message.content);

Best for

Tasks where o4 ranks among LLMDex's top picks.

Compare o4 with…

Frequently asked

  • How much does o4 cost?
    OpenAI has not published per-token API pricing for o4 at the time of writing. Check the official site for current pricing tiers, or compare against alternative models on LLMDex.
  • What is o4's context window?
    o4 supports a context window of 200K tokens.
  • Is o4 open source?
    No. o4 is a closed-weight model, you can use it via OpenAI's API but the model weights are not publicly downloadable.
  • When was o4 released?
    o4 was released on Dec 15, 2025 by OpenAI.
Friday digest

Intelligence, distilled weekly.

One short email every Friday, new model launches, leaderboard moves, and pricing drops. Curated by hand. Free, no spam.