Virtual poolant:free

ANT Free

Zero-cost routing across every free model in the pool.

How this pool routes

ant:free routes only to models that cost nothing, drawing on the free-model pool and its own fallback chain across providers. Throughput is bounded by upstream free-tier rate limits rather than your balance.

Best For

PrototypingSide projectsCost-free experimentation

Where requests went

Live routing over the last 30 days

Requests

15,446

Success rate

99.95%

Avg latency

417ms

gemini-2.5-flashSemantic-cache33.9% · 5,242

deepseek-ai/DeepSeek-V3-0324Semantic-cache26.6% · 4,101

openai/gpt-oss-120bSemantic-cache9.6% · 1,486

command-a-03-2025Semantic-cache9.6% · 1,486

meta-llama/Llama-4-Scout-17B-16E-InstructSemantic-cache5.5% · 843

nvidia/nemotron-3-nano-30b-a3bSemantic-cache4.5% · 696

Gemini 2.5 FlashNVIDIA NIM2.1% · 326

openai/gpt-oss-20bSemantic-cache1.9% · 299

deepseek/deepseek-v3-0324Semantic-cache1.9% · 293

DeepSeek V3.1Hugging Face1.5% · 234

Models in this pool

51 models across 12 providers currently qualify. Per-request limits (context size, cost caps, your model allowlist) narrow this further at routing time.

Showing the 24 highest-scoring of 51.

Quick Start

from openai import OpenAI

client = OpenAI(
    base_url="https://api.antbase.ai/v1",
    api_key="YOUR_ANT_API_KEY",
)

response = client.chat.completions.create(
    model="ant:free",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

Routing weights

Speed50/100

Quality30/100

Cost100/100

Selection criteria

Speed weight50/100

Quality weight30/100

Cost weight100/100

Model IDant:free

Other pools

Every virtual pool accepts the same request shape — swap the model id to change the routing policy.

Browse all virtual models →