BroadRouter
Get started

GPT OSS 20B API pricing

Updated

Among the 10 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest blended standard price for GPT OSS 20B is AkashML (promotional) at USD 0.018 per 1M input tokens and USD 0.09 per 1M output tokens. The median across 10 standard prices is USD 0.04 / USD 0.19. Each price is read from the provider's own pricing page.

OpenAI · slug gpt-oss-20b · up to 131,072 tokens of context · 10 providers with a standard price. Every price is read from the provider's own pricing page and linked in its row, checked 2026-11-15. No aggregator feed. Download all prices (every provider and model, with sources, CSV or JSON). How it compares with the market: the LLM API price index. Archived list prices since 2025-08-06 (19 archived price points, 1 list-price change): LLM API price history.

Does OpenAI sell GPT OSS 20B directly?

OpenAI publishes no price for GPT OSS 20B on its own API that we could find, so this page compares hosts only.

GPT OSS 20B model facts

Facts from OpenAI's model card for openai/gpt-oss-20b and the Hugging Face model API, read 2026-11-16. BroadNet did not verify them, and hosts may serve the model at a different precision. BroadRouter is BroadNet's site.

GPT OSS 20B: facts as stated by the model card and the Hugging Face model API, with sources
FactAs statedSource
Open weightsYes. The weights are in a public Hugging Face repository (openai/gpt-oss-20b) with no access gate.Hugging Face model API, checked 2026-11-16
Quote
“"private": false, "gated": false, 4 .safetensors files listed”
LicenceApache 2.0 (apache-2.0)OpenAI model card metadata, checked 2026-11-16
Quote
“license: apache-2.0”
Total parameters21BOpenAI model card, checked 2026-11-16
Quote
“(21B parameters with 3.6B active parameters)”
Active parameters per token3.6BOpenAI model card, checked 2026-11-16
Quote
“(21B parameters with 3.6B active parameters)”
Context lengthNot found on OpenAI's model card when we read it; see the context column in the price tablechecked 2026-11-16
Input and outputText in, text out (the card's task tag is text-generation)OpenAI model card metadata, checked 2026-11-16
Quote
“pipeline_tag: text-generation”
Tool callingYesOpenAI model card, checked 2026-11-16
Quote
“Use the models’ native capabilities for function calling”
First published on Hugging Face (not necessarily the release date)2025-08-04Hugging Face model API, checked 2026-11-16
Quote
“"createdAt": "2025-08-04T22:33:29.000Z"”

GPT OSS 20B price per provider

USD per 1M tokens. Standard prices are ranked on a blend of 3 input tokens to 1 output token, cheapest first; the dearest standard price is 6.3× the cheapest. Free rows, the maker's serving version and other service tiers follow, tagged and unranked.

GPT OSS 20B API prices by provider
#ProviderInputCached inputOutputBlended 3:1ContextPrecisionSource
1AkashML cheapest promo
openai/gpt-oss-20b
Notes and source quote

Labelled "Limited Trial" on AkashML.

“Price (per 1M Tokens) Input Output Cache Read GPT OSS 20B $0.018 $0.09 $0.009”
USD 0.018USD 0.009USD 0.09USD 0.036131,072not statedAkashML page
checked 2026-11-15
re-checked 2026-11-16
2CoreWeave Forge
openai/gpt-oss-20b
Notes and source quote
“OpenAI GPT OSS 20B $0.03 $0.13 -”
USD 0.03not publishedUSD 0.13USD 0.055131,072not statedCoreWeave Forge page
checked 2026-11-15
re-checked 2026-11-16
3DeepInfra
openai/gpt-oss-20b
Notes and source quote
“$0.03 in $0.14 out / 1M tokens”
USD 0.03not publishedUSD 0.14USD 0.0575131,072BFLOAT16DeepInfra page
checked 2026-11-14
re-checked 2026-11-16
4Featherless
openai/gpt-oss-20b
Notes and source quote

Developer plan (per-token billing, from 50 USD a month in credits). No cached price published.

“PRICING: INPUT $0.04 / OUTPUT $0.15”
USD 0.04not publishedUSD 0.15USD 0.0675131,072FP8Featherless page
checked 2026-11-15
re-checked 2026-11-16
5SiliconFlow
Notes and source quote

Context shown as 131K.

“gpt-oss-20b 131K $0.04 $0.18”
USD 0.04not publishedUSD 0.18USD 0.075131,072not statedSiliconFlow page
checked 2026-11-14
re-checked 2026-11-16
6Parasail
Notes and source quote
“gpt-oss-20b $0.04 $0.20 $0.02”
USD 0.04USD 0.02USD 0.20USD 0.08not statednot statedParasail page
checked 2026-11-14
re-checked 2026-11-16
7Google Vertex AI
gpt-oss-20b
Notes and source quote
“gpt-oss-20b Input $0.07 … Output $0.25 … Cache Hit $0.007”
USD 0.07USD 0.007USD 0.25USD 0.115not statednot statedGoogle Vertex AI page
checked 2026-11-14
re-checked 2026-11-16
8Amazon Bedrock
Notes and source quote

Standard tier pricing table. Shown for region US East (N. Virginia) on the pricing page (values filled from AWS's published price list for us-east-1).

“gpt-oss-20b $0.07 $0.30”
USD 0.07not publishedUSD 0.30USD 0.1275not statednot statedAmazon Bedrock page
checked 2026-11-14
re-checked 2026-11-16
9Groq
openai/gpt-oss-20b
Notes and source quote
“$0.075 input$0.30 output | Cached Input $0.037”
USD 0.075USD 0.037USD 0.30USD 0.1312131,072not statedGroq page
checked 2026-11-14
re-checked 2026-11-16
10Cloudflare Workers AI
@cf/openai/gpt-oss-20b
Notes and source quote

Cloudflare publishes this USD per-token price alongside the neuron price.

“@cf/openai/gpt-oss-20b $0.200 per M input tokens $0.300 per M output tokens”
USD 0.20not publishedUSD 0.30USD 0.225not statednot statedCloudflare Workers AI page
checked 2026-11-14
re-checked 2026-11-16
—Google Vertex AI batch
gpt-oss-20b
Notes and source quote
“gpt-oss-20b Batch Input $0.035 … Batch Output $0.125”
USD 0.035not publishedUSD 0.125USD 0.0575not statednot statedGoogle Vertex AI page
checked 2026-11-14
re-checked 2026-11-16
—Amazon Bedrock batch
Notes and source quote

Batch tier pricing table. Shown for region US East (N. Virginia) on the pricing page (values filled from AWS's published price list for us-east-1).

“gpt-oss-20b $0.035 $0.15”
USD 0.035not publishedUSD 0.15USD 0.06375not statednot statedAmazon Bedrock page
checked 2026-11-14
re-checked 2026-11-16
—DeepInfra flex / spot
openai/gpt-oss-20b
Notes and source quote

Flex service tier (0.8× standard).

“Flex (0.8×) $0.024 $0.112”
USD 0.024not publishedUSD 0.112USD 0.046131,072BFLOAT16DeepInfra page
checked 2026-11-14
re-checked 2026-11-16
—Amazon Bedrock flex / spot
Notes and source quote

Flex tier pricing table. Shown for region US East (N. Virginia) on the pricing page (values filled from AWS's published price list for us-east-1).

“gpt-oss-20b $0.035 $0.15”
USD 0.035not publishedUSD 0.15USD 0.06375not statednot statedAmazon Bedrock page
checked 2026-11-14
re-checked 2026-11-16
—DeepInfra priority
openai/gpt-oss-20b
Notes and source quote

Priority service tier (1.5× standard).

“Priority (1.5×) $0.045 $0.21”
USD 0.045not publishedUSD 0.21USD 0.08625131,072BFLOAT16DeepInfra page
checked 2026-11-14
re-checked 2026-11-16
—Amazon Bedrock priority
Notes and source quote

Priority tier pricing table. Shown for region US East (N. Virginia) on the pricing page (values filled from AWS's published price list for us-east-1).

“gpt-oss-20b $0.1225 $0.525”
USD 0.1225not publishedUSD 0.525USD 0.2231not statednot statedAmazon Bedrock page
checked 2026-11-14
re-checked 2026-11-16

Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).

What does GPT OSS 20B cost for a real workload?

Monthly cost over 30 days at each standard price, using the presets of the LLM API cost calculator. Providers without a cached price bill cached input at the input price.

  • Support chatbot: 2,000 requests a day, 1,500 input tokens (60% cached), 300 output.
  • RAG over documents: 1,000 requests a day, 8,000 input tokens (30% cached), 500 output.
  • Coding agent (one developer): 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning.
Monthly cost of three workloads on GPT OSS 20B, by provider
ProviderSupport chatbotRAG over documentsCoding agent (one developer)
AkashML (promo)USD 2.75USD 5.02USD 26.73
CoreWeave ForgeUSD 5.04USD 9.15USD 65.70
DeepInfraUSD 5.22USD 9.30USD 66.60
FeatherlessUSD 6.30USD 11.85USD 85.50
SiliconFlowUSD 6.84USD 12.30USD 88.20
ParasailUSD 6.12USD 11.16USD 59.40
Google Vertex AIUSD 7.40USD 16.01USD 52.11
Amazon BedrockUSD 11.70USD 21.30USD 153.00
GroqUSD 10.10USD 19.76USD 103.86
Cloudflare Workers AIUSD 23.40USD 52.50USD 387.00

For the support chatbot, the cheapest price (AkashML, USD 2.75 a month) saves USD 3.82 against the median (USD 6.57). For the coding agent the gap is USD 49.32 a month (65% lower).

Which GPT OSS 20B provider is cheapest for your traffic?

AkashML has the lowest input price (USD 0.018) and the lowest output price (USD 0.09), so it is the cheapest standard price at every mix of input and output tokens, before cache discounts.

GPT OSS 20B: price per 1M tokens at 3 input to 1 output, by share of input served from cache
ProviderNo cache hits50% cached90% cached
AkashMLUSD 0.036 cheapestUSD 0.03263 cheapestUSD 0.02993 cheapest
CoreWeave Forge (no cache price)USD 0.055USD 0.055USD 0.055
DeepInfra (no cache price)USD 0.0575USD 0.0575USD 0.0575
Featherless (no cache price)USD 0.0675USD 0.0675USD 0.0675
SiliconFlow (no cache price)USD 0.075USD 0.075USD 0.075
ParasailUSD 0.08USD 0.0725USD 0.0665

USD per 1M tokens, 3 input tokens to 1 output token, each provider's cheapest standard row; 4 dearer providers are in the price table above. A provider without a cached price bills cached input at the input price. Any charge for writing to the cache, which some providers add, is not included.

4 of 10 providers publish a cached-input price below the input price: AkashML (50% off), Parasail (50% off), Google Vertex AI (90% off), Groq (51% off). CoreWeave Forge, DeepInfra, Featherless, SiliconFlow, Amazon Bedrock and Cloudflare Workers AI publish none, so cached tokens there bill at the input price.

AkashML, CoreWeave Forge, DeepInfra, Featherless, SiliconFlow and Groq state a context window of 131,072 tokens; Parasail, Google Vertex AI, Amazon Bedrock and Cloudflare Workers AI do not state one.

What does GPT OSS 20B cost in other languages?

The same text needs a different number of tokens in each language. With the gpt-oss (o200k_harmony) tokenizer, sending 1M characters costs this much at the cheapest standard input price we found for GPT OSS 20B:

ContentInput per 1M charactersCheapest provider
EnglishUSD 0.0064CoreWeave Forge
Arabic MSAUSD 0.0096CoreWeave Forge
HindiUSD 0.0097CoreWeave Forge
ChineseUSD 0.023CoreWeave Forge
CodeUSD 0.0076CoreWeave Forge

From BroadRouter's real cost by language dataset, which covers 15 languages and content types for every model we could match to a measured tokenizer.

Published benchmark results

BroadNet did not run these benchmarks; results are as published by Epoch AI and may use different settings. BroadRouter is BroadNet's site.

Published benchmark results for GPT OSS 20B, with sources
BenchmarkResultSetting or noteSource
GPQA Diamond0.460High reasoning setting; Epoch run gpt-oss-20b_high, mean score on a 0 to 1 scaleEpoch AI, checked 2026-11-16

Epoch AI rows: Epoch AI, Capabilities & benchmarking, licensed CC BY 4.0, shown as the mean score on a 0 to 1 scale, rounded to three decimals (hover a value for the exact figure in Epoch's data file).

Estimate your own bill

The OpenAI models whose cheapest standard price is closest to GPT OSS 20B's (USD 0.036 blended):

OpenAI models priced closest to GPT OSS 20B
ModelCheapest inputCheapest outputMax contextProviders
OpenAI GPT OSS 120BUSD 0.03USD 0.17131,07221
GPT OSS Safeguard 20BUSD 0.07USD 0.20131,0723
GPT-5 NanoUSD 0.05USD 0.40—3
GPT-6 LunaUSD 0.10USD 0.501,050,0006
GPT-4o miniUSD 0.15USD 0.60—3
GPT-5.6 LunaUSD 0.20USD 1.201,050,0006

Priced closest from other makers: Mistral-Nemo-Instruct-2407 (Mistral AI, USD 0.02925) · Gemma 3 4B (Google, USD 0.05) · Llama 3 8B Stheno V3.2 (Sao10K, USD 0.05).

Compare GPT OSS 20B with

Also compared: vs OpenAI GPT OSS 120B · vs Gemma 4 26B A4B · vs Qwen3.6 35B A3B.

Every GPT OSS 20B price comparison · All model-vs-model price comparisons

All 428 models: API prices compared · Cheapest LLM APIs right now · LLM API price index (open data) · Download all prices (CSV, JSON) · Price change log

Questions

How much does the GPT OSS 20B API cost per token?

From USD 0.018 per 1M input tokens (USD 0.000000018 per token) and USD 0.09 per 1M output tokens (USD 0.00000009 per token) at the lowest standard prices we found. The median of 10 standard prices is USD 0.04 input and USD 0.19 output. Each price was read from the provider's own pricing page; the source is linked in its row.

Which provider is cheapest for GPT OSS 20B?

Among the 10 providers we checked (prices read 2026-11-14 to 2026-11-15), the lowest is AkashML (promotional price), ranked on a blend of 3 input tokens to 1 output token (USD 0.036 per 1M). It also has the lowest input and the lowest output price. Batch, flex, priority and fast tiers are shown in the table but not ranked.

Does OpenAI sell GPT OSS 20B on its own API?

OpenAI publishes no price for GPT OSS 20B on its own API that we could find, so there is no maker price to compare. The lowest-priced host we checked is AkashML at USD 0.018 / USD 0.09 per 1M tokens.

Does GPT OSS 20B have cached input pricing?

4 of 10 standard prices include one. The lowest is USD 0.007 per 1M at Google Vertex AI; the median is USD 0.0145. Providers without one bill cached input at the input price. What cached input means.

What is the GPT OSS 20B context window?

131,072 tokens on every provider that states it.

How much does a coding agent cost per month on GPT OSS 20B?

For 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning, over 30 days: USD 26.73 a month at AkashML, USD 76.05 at the median price and USD 387.00 at Cloudflare Workers AI. Try your own numbers.

Is there a batch, flex or priority price for GPT OSS 20B?

Yes: batch from Google Vertex AI at USD 0.035 / USD 0.125; flex / spot from DeepInfra at USD 0.024 / USD 0.112; priority from DeepInfra at USD 0.045 / USD 0.21 per 1M input / output tokens. These tiers trade speed or availability for price (batch runs within hours, flex and spot can be slow or refused at capacity, priority and fast cost more for speed or reliability), so they are shown but not ranked.

Is GPT OSS 20B free to use through an API?

No provider in our data lists a free price for GPT OSS 20B as of 2026-11-15. Free tiers and credits by provider (checked 2026-11-14) lists general offers that do not name a model; all free LLM API models are on one page.

Is GPT OSS 20B open source, with open weights, and what is its licence?

OpenAI publishes the weights on Hugging Face as openai/gpt-oss-20b, in a public repository with no access gate. The licence, as the OpenAI model card metadata states it: Apache 2.0 (apache-2.0). Open weights means you can download and run the model yourself; what you may do with it, including commercial use, is set by that licence, so read it in full (checked 2026-11-16). These facts come from OpenAI's card and the Hugging Face model API; BroadNet did not verify them, and hosts may serve the model at a different precision.

How many parameters does GPT OSS 20B have?

Total parameters, as the OpenAI model card states them: 21B. Active parameters per token: 3.6B. (checked 2026-11-16; BroadNet did not verify these figures.)

Does GPT OSS 20B support tool calling?

Yes, according to the OpenAI model card: “Use the models’ native capabilities for function calling” (checked 2026-11-16). Whether the tools parameter works through an API also depends on the provider serving it, so check that provider's docs. BroadNet did not test it.

How do I call GPT OSS 20B with the OpenAI SDK?

BroadNet does not serve GPT OSS 20B. Use the OpenAI-compatible base URL of a provider that lists it: create the client with OpenAI(base_url=..., api_key=...) and pass that provider's model ID. Documented base URLs for providers on this page, in alphabetical order: DeepInfra (https://api.deepinfra.com/v1/openai, model openai/gpt-oss-20b) and Groq (https://api.groq.com/openai/v1, model openai/gpt-oss-20b). Each provider page quotes the provider's docs with a check date.

How does GPT OSS 20B score on benchmarks?

As published, Epoch AI measured 0.460 on GPQA Diamond in its own run, as a mean score on a 0 to 1 scale (checked 2026-11-16). BroadNet did not run these benchmarks, and sources may use different settings, so the figures are not directly comparable. Every published result we found, with its setting and source.

Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity