Grok 4.20 Multi-Agent pricing for bug detection

Quick start

Call Grok 4.20 Multi-Agent with the OpenAI Python SDK pointed at api.x.ai:

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.x.ai/v1",
    api_key=os.environ["XAI_API_KEY"],
)
response = client.chat.completions.create(
    model="grok-4.20-multi-agent-0309",
    max_tokens=2000,
    messages=[{"role": "user", "content": "Your prompt here"}]
)
# Cost: ~$0.0112 per call at standard rates
# Input:  $1.25/MTok
# Output: $2.50/MTok

How Grok 4.20 Multi-Agent compares on cost

Per-call cost for all models scoring 50+ on bug detection benchmarks. Hover for details.

Cheapest viable option: Claude Sonnet 5 at $0.0300/call (-167% less, 96/100 quality).

What drives bug detection cost

Bug detection sends a file and gets back a short list of suspected defects, so input dominates and output stays small. That makes it cheap per call and tempting to run continuously, which is the real cost driver: frequency, not per-call price. Caching the unchanged parts of the codebase and running only on changed files is worth more than switching models.

For Grok 4.20 Multi-Agent specifically, that means output tokens cost 2.0x more than input, at $2.50 vs $1.25 per MTok. At this use case’s 5,000/2,000 token split, output accounts for 44% of the $0.0112 per-call cost.

Grok 4.20 Multi-Agent quality benchmarks for bug detection

Benchmark data for bug detection is pending for Grok 4.20 Multi-Agent. Test against your own evaluation suite before moving to production.

Performance across all task types

Grok 4.20 Multi-Agent scores by task type. Bug Detection highlighted.

Confidence scores are a third-party benchmark snapshot from week 29 of 2026. This snapshot is 10 weeks old, so treat the scores as directional and validate against your own evals.

About Grok 4.20 Multi-Agent

Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.

Other models from Grok

You can compare Grok 4.20 Multi-Agent head-to-head against any other model on our model comparison pages, or track how pricing has changed over time on the pricing history tracker.

Frequently asked questions