Virtual poolant:balanced

ANT Balanced

Even speed and quality, with a quality floor of 60.

Try in Playground
Quality floor: 60/100 or above

How this pool routes

ant:balanced is ant:auto with a guard rail: models scoring below 60 are excluded outright. Reach for it when you want sensible defaults but cannot tolerate the bottom of the catalog.

Best For

Production chatSummarizationGeneral assistants

Where requests went

Live routing over the last 30 days

Requests

317

Success rate

0.63%

Avg latency

106ms

ant:balancedUnknown99.1% · 314
deepseek/deepseek-v4.1-flashOpenRouter0.3% · 1
openai/o3-miniOpenRouter0.3% · 1
qwen/qwen3-next-80b-a3b-instructOpenRouter0.3% · 1

Models in this pool

402 models across 16 providers currently qualify. Per-request limits (context size, cost caps, your model allowlist) narrow this further at routing time.

o3

OpenAI200K👁🧠Q:96

Claude Sonnet 5

Anthropic1M👁🧠Q:96

o3

OpenAI200K👁🧠Q:96

Anthropic: Claude Opus 4.5

OpenRouter200K👁Q:95

Anthropic: Claude Opus 5 (batch)

OpenRouter1M👁Q:95

Anthropic: Claude Opus 4.6

OpenRouter1M👁Q:95

Anthropic: Claude Opus 4.7

OpenRouter1M👁Q:95

Anthropic: Claude Opus 5.5

OpenRouter1M👁Q:95

Anthropic: Claude Opus Latest

OpenRouter1M👁Q:95

GPT-5

OpenAI400K👁🧠Q:95

Anthropic: Claude Opus 4.1

OpenRouter200K👁Q:95

Anthropic: Claude Opus 4.8

OpenRouter1M👁Q:95

Anthropic: Claude Opus 5

OpenRouter1M👁Q:95

Anthropic: Claude Opus 5.5 (batch)

OpenRouter1M👁Q:95

Claude Sonnet 4

Anthropic200K👁🧠Q:92

GPT-4.1

OpenAI1.0M👁Q:88

Qwen 3 235B

Alibaba131K🧠Q:88

Sonar Deep Research

Perplexity127K🧠Q:88

GPT-4.1

OpenAI1.0M👁Q:88

OpenAI: GPT-4

OpenRouter8KQ:88

DeepSeek V4 Pro

DeepSeek1M🧠Q:88

Claude 3.5 Sonnet

Anthropic200K👁Q:88

Command A (03-2025)

Cohere256K🧠Q:88

DeepSeek R1 671B (Venice)

Venice66K🧠Q:88

Showing the 24 highest-scoring of 402.

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:balanced",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

Routing weights

Speed50/100
Quality60/100
Cost50/100

Selection criteria

Quality floorrequired60/100 or above
Speed weight50/100
Quality weight60/100
Cost weight50/100
Model IDant:balanced

Other pools

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

Browse all virtual models →