Pular para o conteúdo
Alfa
  • Modelos
  • Classificações
  • Ori
Falar com vendas

Rodapé

OpenRouterOpenRouter
© 2026 OpenRouter, Inc

Produto

  • Chat
  • Classificações
  • Avaliações de desempenho
  • Aplicativos
  • Descobrir
  • Modelos
  • Ori
  • Coleções
  • Provedores
  • Ferramentas
  • Preços
  • Business
  • Enterprise
  • Laboratório

Empresa

  • Sobre
  • Blog
  • Clientes
  • Carreiras
    Vagas
  • Privacidade
  • Termos de serviço
  • Central de confiança
  • Suporte
  • Funciona com OR
  • Dados
  • Marca

Desenvolvedor

  • Documentação
  • Referência da API
  • Plataforma para desenvolvedores
  • Status
  • Mapa do site para IA

Conectar

  • Discord
  • GitHub
  • LinkedIn
  • X
  • YouTube
Favicon de openai

OpenAI: gpt-oss-20b

openai/gpt-oss-20b

Pesos do modelo
Comparar

gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI’s Harmony response format and supports reasoning level configuration, fine-tuning, and agentic capabilities including function calling, tool use, and structured outputs.

Modalidades
Preço de entrada / saída
US$ 0,006 / US$ 0,026por 1M
Contexto
131K
Lançado
5 de ago. de 2025
Data limite de conhecimento
jun. de 2024
Comparar
ProvedoresPreçosDesempenhoDisponibilidadeAvaliações de desempenhoAplicativosAtividadePerguntas frequentesExplorar

Provedores

Diferentes empresas hospedam o mesmo modelo. O OpenRouter encaminha sua solicitação para uma delas com base no modo de roteamento que você escolhe — Equilibrado (preço + velocidade), Nitro (mais rápido) ou Floor (mais barato).

Preços

O preço médio que os clientes realmente pagam por este modelo, ao lado dos preços publicados pelos provedores. O cache e os descontos fazem com que o preço efetivamente pago seja frequentemente bem inferior ao preço anunciado.

Desempenho

A vazão indica a velocidade com que o modelo escreve (tokens por segundo — quanto maior, melhor). A latência é o tempo total de ida e volta (quanto menor, melhor). TTFT é o tempo até o primeiro token — quanto tempo leva para você ver algo aparecer (quanto menor, melhor).

Disponibilidade

A disponibilidade é a porcentagem dos últimos 3 dias em que pelo menos um provedor estava respondendo às solicitações. A disponibilidade é a porcentagem de tempo em que a inferência foi realizada com sucesso. O OpenRouter monitora continuamente e usa o próximo melhor provedor quando um deles retorna um erro.

Avaliações de desempenho

Pontuações em avaliações padronizadas e as próprias saídas deste modelo para prompts compartilhados de testes de desempenho. Porcentagens maiores são melhores — e cada posição mostra onde este modelo fica entre os modelos avaliados naquela fonte.

Resumo das pontuações de avaliações comparativas para OpenAI: gpt-oss-20b (Artificial Analysis, Design Arena e Vals AI)
FonteAvaliação comparativaPontuação
Artificial Analysisgpt-oss-20b (High) Coding Index20,7
Artificial Analysisgpt-oss-20b (High) Agentic Index1,2
Artificial Analysisgpt-oss-20b (High) GPQA Diamond68.8%
Artificial Analysisgpt-oss-20b (High) HLE11.0%
Artificial Analysisgpt-oss-20b (High) IFBench65.1%
Artificial Analysisgpt-oss-20b (High) τ²-Bench Telecom60.2%
Artificial Analysisgpt-oss-20b (High) AA-LCR34.7%
Artificial Analysisgpt-oss-20b (High) τ-Bench Banking7.0%
Artificial Analysisgpt-oss-20b (High) GDPval-AA0.0%
Artificial Analysisgpt-oss-20b (High) CritPt1.4%
Artificial Analysisgpt-oss-20b (High) LiveCodeBench77.7%
Artificial Analysisgpt-oss-20b (High) SciCode38.9%
Artificial Analysisgpt-oss-20b (High) Terminal-Bench Hard10.6%
Artificial Analysisgpt-oss-20b (High) Terminal-Bench 4.00.0%
Artificial Analysisgpt-oss-20b (High) Terminal-Bench 2.113.9%
Artificial Analysisgpt-oss-20b (High) MMLU-Pro74.8%
Artificial Analysisgpt-oss-20b (High) AIME 202589.3%
Artificial Analysisgpt-oss-20b (Low) GPQA Diamond61.1%
Artificial Analysisgpt-oss-20b (Low) HLE5.3%
Artificial Analysisgpt-oss-20b (Low) IFBench57.8%
Artificial Analysisgpt-oss-20b (Low) τ²-Bench Telecom50.3%
Artificial Analysisgpt-oss-20b (Low) AA-LCR31.0%
Artificial Analysisgpt-oss-20b (Low) CritPt0.0%
Artificial Analysisgpt-oss-20b (Low) LiveCodeBench65.2%
Artificial Analysisgpt-oss-20b (Low) Terminal-Bench Hard4.5%
Artificial Analysisgpt-oss-20b (Low) MMLU-Pro71.8%
Artificial Analysisgpt-oss-20b (Low) AIME 202562.3%
Design ArenaGPT OSS 20B Models Arena Data Visualization Elo933
Design ArenaGPT OSS 20B Models Arena Website Elo852
Vals AIGPT OSS 20B AIME86.0%
Vals AIGPT OSS 20B CaseLaw v243.8%
Vals AIGPT OSS 20B CorpFin v253.1%
Vals AIGPT OSS 20B GPQA Diamond68.9%
Vals AIGPT OSS 20B LiveCodeBench80.4%
Vals AIGPT OSS 20B LegalBench70.8%
Vals AIGPT OSS 20B MATH 50094.2%
Vals AIGPT OSS 20B MedQA82.9%
Vals AIGPT OSS 20B MGSM89.0%
Vals AIGPT OSS 20B MMLU Pro71.6%
Vals AIGPT OSS 20B TaxEval v263.7%

Aplicativos

Aplicativos públicos que enviam mais tráfego para este modelo. Um bom indicador de como são as cargas de trabalho reais em produção — e uma pista sobre os casos de uso para os quais este modelo é mais adequado.

Atividade

Volume de tokens e tráfego de requisições para este modelo ao longo do tempo.

API

Código pronto para chamar este modelo. A API do OpenRouter é compatível com a OpenAI — a maioria dos SDKs funciona apenas trocando a URL base. A única coisa que muda entre os modelos é o slug do modelo abaixo.

Disponível por meio do roteamento intrarregional da UE

OpenAI: gpt-oss-20b está disponível por meio do roteamento regional na UE. As solicitações enviadas para eu.openrouter.ai são descriptografadas na UE e roteadas apenas para endpoints compartilhados de provedores na Europa (UE/EEE, Reino Unido e Suíça).

Provedores na Europa
Amazon Bedrock
Preços na Europa
  • Amazon Bedrock: US$ 0,07/M tokens de entrada, US$ 0,15/M tokens de saída
URL base
https://eu.openrouter.ai/api/v1
Plano
Requer Business ou Enterprise. Autoatendimento no Business: sem contrato, sem mínimo mensal. OpenRouter Business para roteamento dentro da região da UE

Falha em modo fechado: se nenhum endpoint de provedor compartilhado na Europa estiver disponível, a solicitação retorna um erro em vez de sair da região. As solicitações atendidas pela sua própria chave BYOK vão para a região da sua implantação, que você controla. Documentação de roteamento na região

Explorar mais modelos

Classificações de modelos de IAClassificação
Latência / taxa de transferência
US$ 0,008US$ 0,034US$ 0,00700,46s112 tps
99,42%
US$ 0,018US$ 0,090US$ 0,00901,60s54 tps
99,97%
US$ 0,020US$ 0,100--0,43s128 tps
100,00%
US$ 0,029US$ 0,140US$ 0,02900,55s57 tps
100,00%
US$ 0,030US$ 0,130US$ 0,03000,30s124 tps
99,86%
US$ 0,030US$ 0,140--0,78s66 tps
99,79%
US$ 0,030US$ 0,150US$ 0,02001,80s97 tps
99,97%
US$ 0,040US$ 0,180--1,06s41 tps
89,72%
US$ 0,070US$ 0,150--0,41s415 tps
100,00%
US$ 0,075US$ 0,300US$ 0,03750,38s376 tps
100,00%
Não utilizado no roteamento Standard:Por que estes endpoints não são usados
US$ 0,070US$ 0,150--0,36s126 tps
100,00%
Taxa de transferência
415tok/s

P50, melhor entre os provedores

Latência
0,30s

P50, melhor provedor

Avaliações de desempenho do AutoExacto
GPQA DiamondTAU-BenchDekaLLM68,5%--SiliconFlow63,0%--auto-routing66,7%52,5%Novita58,1%--Darkbloom64,7%51,1%
+8 provedores adicionais
Disponibilidade (3d)O modelo estava acessível. Requisição roteada para um provedor.
100,00%
Disponibilidade (3d)O modelo retornou inferência de qualquer provedor. Erros e respostas vazias contam negativamente para essa métrica.
97,37%

Disponibilidade nos últimos 3 dias

Últimas 72 horas
Disponibilidade 97,37%
Há 3 diasHá 2 diasOntemAgora

Disponibilidade nas últimas 24 horas

OpenRouter Availability
97,10%
Without Routing
93,94%

Quando ocorre um erro em um provedor upstream, podemos nos recuperar roteando para outro provedor saudável, se os filtros da sua solicitação permitirem. Você pode acessar os dados de disponibilidade por provedor de forma programática pela API de endpoints. Saiba mais sobre nossas opções de balanceamento de carga e roteamento de provedores.

1.
Favicon de https://heyditto.ai/
HeyDitto
novo
137Btokens
2.
Favicon de https://jobleads.com/
JobLeads LLM
Use AI to find jobs for you
31,2Btokens
3.
Favicon de https://centralcommand.io/
Central Command
novo
22,8Btokens
4.
Favicon de https://fvchat.app/
FVChat
novo
9,8Btokens
5.
Favicon de https://gdevelop.io/
GDevelop
AI assisted game engine.
7,11Btokens

Perguntas frequentes

gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for lower-latency inference and deployability on consumer or single-GPU hardware.

gpt-oss-20b custa US$ 0,006/M de tokens de entrada e US$ 0,026/M de tokens de saída, com taxas separadas para Cache Read por US$ 0,005/M tokens.

gpt-oss-20b tem uma janela de contexto de 131.072 tokens. Ele aceita até 32.768 tokens de conclusão.

Sim. gpt-oss-20b aceita tools e tool_choice para chamadas de função em 9 dos 10 provedores que o disponibilizam, e as solicitações que enviam ferramentas são roteadas para esses provedores. Também oferece suporte a saídas estruturadas por meio de um esquema JSON em response_format.

gpt-oss-20b é disponibilizado por 10 provedores no OpenRouter: Prime Intellect, Darkbloom, AkashML, DekaLLM, CoreWeave, DeepInfra, Parasail, SiliconFlow e mais 2. As solicitações são roteadas para o melhor provedor disponível, com alternância automática para os demais em caso de falha, e você pode fixar ou excluir provedores com o roteamento de provedores.

gpt-oss-20b foi lançado em 5 de agosto de 2025. Seu conhecimento vai até 30 de junho de 2024.

Mais modelos de OpenAI

GPT-6.1 Terra Pro

GPT-6.1 Terra Pro is the same underlying model as GPT-6.1 Terra, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6.1 Terra and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6.1 Terra instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 1 / US$ 5
GPT-6.1 Terra Pro (batch)

GPT-6.1 Terra Pro is the same underlying model as GPT-6.1 Terra, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6.1 Terra and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6.1 Terra instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 0,50 / US$ 2,50
GPT-6.1 Terra

GPT-6.1 Terra is the mid-tier model in OpenAI's GPT-6.1 series, priced below GPT-6.1 Sol. It is suited for everyday agentic, coding, and writing work, is particularly strong at understanding screenshots and on-screen content, and is efficient enough for high-volume workloads.

TextoContexto de 1,1MUS$ 1 / US$ 5
GPT-6.1 Terra (batch)

GPT-6.1 Terra is the mid-tier model in OpenAI's GPT-6.1 series, priced below GPT-6.1 Sol. It is suited for everyday agentic, coding, and writing work, is particularly strong at understanding screenshots and on-screen content, and is efficient enough for high-volume workloads.

TextoContexto de 1,1MUS$ 0,50 / US$ 2,50
GPT-6 Luna Decisions

GPT-6 Luna Decisions is GPT-6 Luna served through OpenAI's Decisions API. Instead of generating text, it reads the content passed as state (text, JSON, or images) and returns typed, probabilistic answers to named questions in a single request: the probability of yes for a yes/no question (noul), a probability for every option plus the most likely one (choice), or a probability for every level of an ordered rubric plus the expected score (score).

It is built for classification, routing, moderation, and rubric grading where application code thresholds the returned numbers rather than parsing a chat reply. A request can carry up to 200 questions about the same content.

DecisõesUS$ 0,10 / US$ 0
GPT-6.1 Sol Pro

GPT-6.1 Sol Pro is the same underlying model as GPT-6.1 Sol, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6.1 Sol and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6.1 Sol instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 2 / US$ 10
GPT-6.1 Sol Pro (batch)

GPT-6.1 Sol Pro is the same underlying model as GPT-6.1 Sol, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6.1 Sol and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6.1 Sol instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 1 / US$ 5
GPT-6.1 Sol

GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional work, and multi-step business workflow automation, and approaches Astra-level results on these tasks at a much lower cost. Compared with GPT-6 Sol, it makes fewer factual errors and is more reliable at respecting explicit restrictions and user intent during agentic tasks.

TextoContexto de 1,1MUS$ 2 / US$ 10
GPT-6.1 Sol (batch)

GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional work, and multi-step business workflow automation, and approaches Astra-level results on these tasks at a much lower cost. Compared with GPT-6 Sol, it makes fewer factual errors and is more reliable at respecting explicit restrictions and user intent during agentic tasks.

TextoContexto de 1,1MUS$ 1 / US$ 5
GPT-6 Luna Pro

GPT-6 Luna Pro is the same underlying model as GPT-6 Luna, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6 Luna and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6 Luna instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 0,10 / US$ 0,50
GPT-6 Luna Pro (batch)

GPT-6 Luna Pro is the same underlying model as GPT-6 Luna, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6 Luna and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6 Luna instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 0,05 / US$ 0,25
GPT-6 Luna

GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic tasks, and at higher reasoning effort it can take on complex software engineering and computer-use tasks that previously called for a Sol-tier model. It shares the GPT-6 family's gains in factual reliability and its clearer, more concise communication style.

TextoContexto de 1,1MUS$ 0,10 / US$ 0,50
GPT-6 Luna (batch)

GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic tasks, and at higher reasoning effort it can take on complex software engineering and computer-use tasks that previously called for a Sol-tier model. It shares the GPT-6 family's gains in factual reliability and its clearer, more concise communication style.

TextoContexto de 1,1MUS$ 0,05 / US$ 0,25
GPT-6 Sol Pro

GPT-6 Sol Pro is the same underlying model as GPT-6 Sol, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6 Sol and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6 Sol instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 2 / US$ 10
GPT-6 Sol Pro (batch)

GPT-6 Sol Pro is the same underlying model as GPT-6 Sol, served with reasoning.mode set to pro for higher-quality responses on complex tasks.

Cost note: pro mode spends far more reasoning tokens per request, so a typical request costs several times more than the same request on GPT-6 Sol and takes much longer to complete. It is intended for hard, high-stakes problems where the extra accuracy justifies the cost. For everyday coding, agentic, and chat workloads, use GPT-6 Sol instead.

Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode

TextoContexto de 1,1MUS$ 1 / US$ 5
GPT-6 Sol

GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional work, agentic coding, business workflow automation, and computer use, and is particularly strong at long-horizon software engineering tasks in real codebases. It approaches Astra-level factual reliability at a much lower cost and shares Astra's clearer, more concise communication style in technical and coding conversations.

TextoContexto de 1,1MUS$ 2 / US$ 10
GPT-6 Sol (batch)

GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional work, agentic coding, business workflow automation, and computer use, and is particularly strong at long-horizon software engineering tasks in real codebases. It approaches Astra-level factual reliability at a much lower cost and shares Astra's clearer, more concise communication style in technical and coding conversations.

TextoContexto de 1,1MUS$ 1 / US$ 5
GPT Astra Latest

This model always redirects to the latest model in the GPT Astra family.

TextoContexto de 1,1M
GPT Sol Latest

This model always redirects to the latest model in the GPT Sol family.

TextoContexto de 1,1M
GPT Terra Latest

This model always redirects to the latest model in the GPT Terra family.

TextoContexto de 1,1M