gpt-oss-120b free API: where to use it for free

Where gpt-oss-120b is free

OpenAI publishes the weights under Apache 2.0, and three providers serve it on a free tier; Groq now recommends it in place of the Llama models it retired from free use.

Listed, but not free

• Cerebras (gpt-oss-120b): trial credits only; Cerebras has no permanent free tier.
• OpenRouter (openai/gpt-oss-120b): paid only; there is no :free variant.

What gpt-oss-120b is

gpt-oss-120b is OpenAI's open-weight reasoning model, a 117B mixture-of-experts with 5.1B active, with adjustable reasoning effort, function calling and structured outputs. gpt-oss-20b is the 21B sibling.

Parameters

117B total, 5.1B active (MoE)

Context window

131,072 tokens

Input

Text

Output

Text

Reasoning

Low, medium or high effort

Tools

Function calling and structured outputs

Before you rely on it

• It is the oldest model on this list, released in August 2025, but still one of the most widely free.
• Groq's free plan caps tokens at 8,000 per minute, which limits long prompts.

Use gpt-oss-120b with automatic failover

A single free tier runs out. Ficelle puts the free models of every provider you configure behind one local OpenAI-compatible endpoint. Call a routing profile and gpt-oss-120b is one of the models it can pick; when it rate-limits or fails, the next eligible free model answers. To pin gpt-oss-120b alone, call the id that ficelle models lists for it, knowing a pinned model has no failover.

from openai import OpenAI

client = OpenAI(base_url="http://127.0.0.1:8646/v1", api_key="ficelle-local")

# A routing profile keeps failover: gpt-oss-120b is one of the free
# models it can pick, and the next eligible one answers if it fails.
client.chat.completions.create(
    model="ficelle/auto-reasoning",
    messages=[{"role": "user", "content": "..."}],
)

Routing profiles that fit what gpt-oss-120b is built for:

• Reasoning ficelle/auto-reasoning
• Tool calling ficelle/auto-tools
• JSON & structured output ficelle/auto-json

Whether a profile routes to gpt-oss-120b on your install depends on your providers' live catalogs and Ficelle's eligibility checks, decided at request time.

Questions