Home/Compare/Models
Source-tracked model comparison

GPT-OSS 20B vs 120B on Groq: which costs less for agent loops?

Compare verified pricing and model limits for choosing between two GPT-OSS sizes on the same GroqCloud API. The cost example uses a ai agent loops workload and does not assume equal model quality.

Direct cost answer: GPT-OSS 20B on Groq is estimated at $34.63 per month and GPT-OSS 120B on Groq at $69.30 for 6,000 input tokens, 2,500 output tokens, and 30,000 monthly requests. GPT-OSS 20B on Groq is $34.67 lower under these assumptions. This does not identify a quality winner.
Same-series comparisonAI agent loopsSources checked 2026-07-15 / 2026-07-15
Adjust this comparisonOpen workload calculator
Pair-specific guidance

Route GPT-OSS work by task difficulty

Both API identifiers use the same tracked context and maximum output limits on GroqCloud, while the 120B token rates are twice the 20B rates. Treat quality and retry behavior as test results, not assumptions from model size.

Decision factorGPT-OSS 20B on GroqGPT-OSS 120B on Groq
Model creatorOpenAIOpenAI
GroqCloud token price$0.075 input and $0.30 output per million$0.15 input and $0.60 output per million
Tracked limits131,072 context and 65,536 maximum output131,072 context and 65,536 maximum output
StackLens assessment

Rollout checks for this pair

  • Test both models on the same difficult agent traces before setting a default route.
  • Count retries, tool calls, and output tokens when measuring cost per completed task.
  • Consider an escalation rule only after the 20B failure cases are clearly defined.
Tracked facts

Pricing and model limits

Prices are USD per 1M tokens under each model's verified default profile.

FieldGPT-OSS 20B on GroqGPT-OSS 120B on Groq
API access providerGroqCloudGroqCloud
Model creatorOpenAIOpenAI
API model IDopenai/gpt-oss-20bopenai/gpt-oss-120b
Input / 1M$0.075$0.15
Cached input / 1M$0.037$0.075
Output / 1M$0.30$0.60
Context window131,072 tokens131,072 tokens
Maximum output65,536 tokens65,536 tokens
Accepted inputtexttext
Example workload

AI agent loops cost scenario

6,000 input and 2,500 output tokens per request, 30,000 monthly requests, and 20% cached input.

Groq

GPT-OSS 20B on Groq

$34.63 / month
Input cost
$12.13
Output cost
$22.50
Per 1,000 calls
$1.1544
Pricing profile
standard
View model details
Groq

GPT-OSS 120B on Groq

$69.30 / month
Input cost
$24.30
Output cost
$45.00
Per 1,000 calls
$2.31
Pricing profile
standard
View model details

Cost result: GPT-OSS 20B on Groq is $34.67 lower per month for these assumptions. This is a price comparison, not a model-quality ranking.

StackLens assessment

Questions to answer before choosing

  • Which representative agent tasks need the larger model?
  • How do retries and generated reasoning tokens affect completed-task cost?
  • Can routine requests route to 20B while selected failures escalate to 120B?
Workload caveat

What this estimate leaves out

Treat monthly requests as model calls, not user tasks. Tool execution and external API charges are excluded.

Latency, reliability, output quality, retries, regional processing, and provider-specific tool charges can change the practical decision.

Continue comparing

Explore GPT-OSS 20B on GroqExplore GPT-OSS 120B on GroqBrowse all modelsCompare other models
Related active comparisons

Continue with a nearby model decision

These comparisons share a model or provider decision with this page.

cost

GPT-OSS 120B on Groq vs deepseek-v4-pro

GPT-OSS 120B on Groq vs DeepSeek V4 Pro: which has lower coding API cost?

Compare API cost