Models
The models Standards offers, published from a Standards workspace.
Conversation models
Auto is the routing mode that picks a model per turn. A chip names a task the model is qualified for and the highest demand it serves — light, standard or heavy, each a ceiling that includes the levels below.
- GPT-6 LunaDefaultRecommendedopenai/gpt-6-lunaAutoconversation · standardwriting · standardcoding · standardanalysis · standardextraction · standard1.1M$0.10 / $0.50
- GPT-6 Luna Fastopenai/gpt-6-luna-fastAutoconversation · lightwriting · lightcoding · lightanalysis · lightextraction · light1.1M$0.20 / $1
- GPT-6 Solopenai/gpt-6-solAutoconversation · heavywriting · heavycoding · heavyanalysis · heavyextraction · standard1.1M$2 / $10
- GPT-6 Sol Fastopenai/gpt-6-sol-fastAutoconversation · standardwriting · standardcoding · heavyanalysis · heavyextraction · standard1.1M$4 / $20
- Claude Haiku 4.5anthropic/claude-haiku-4.5Autoconversation · lightwriting · lightcoding · lightextraction · light200K$1 / $5
- Claude Sonnet 5Recommendedanthropic/claude-sonnet-5Autoconversation · standardwriting · standardcoding · lightanalysis · standardextraction · standard1M$2 / $10
- Claude Opus 5.5anthropic/claude-opus-5.5Autoconversation · heavywriting · heavycoding · heavyanalysis · heavyextraction · standard1M$4 / $20
- Gemini 3.5 Flash LiteUtilitygoogle/gemini-3.5-flash-liteAutoconversation · lightcoding · lightanalysis · light1M$0.30 / $2.50
- Gemini 3.8 FlashRecommendedgoogle/gemini-3.8-flashAutoconversation · standardcoding · heavyanalysis · heavy1M$0.75 / $3.75
- Grok 4.6spacexai/grok-4.6Autoconversation · lightcoding · heavyanalysis · heavy500K$2 / $6
- DeepSeek V4.1 FlashRecommendeddeepseek/deepseek-v4.1-flashAutocoding · heavyanalysis · standard1.0M$0.30 / $1.20
- DeepSeek V4 Prodeepseek/deepseek-v4-proAutoconversation · lightwriting · lightcoding · heavyanalysis · heavyextraction · standard1M$0.66 / $1.98
- Kimi K3moonshotai/kimi-k3Autoconversation · standardwriting · standardcoding · heavyanalysis · heavy1M$3 / $15
- Qwen 3.8 Flashalibaba/qwen3.8-flashAutomanual only991K$0.15 / $0.47
- Qwen 3.8 Maxalibaba/qwen3.8-max-0902Autoconversation · standardwriting · lightcoding · heavyanalysis · heavyextraction · standard991K$2 / $6
- Ministral 14Bmistral/ministral-14bAutomanual only262.1K$0.20 / $0.20
- Mistral Medium 3.5mistral/mistral-medium-3.5Autoconversation · lightextraction · light262.1K$1.50 / $7.50
- Mistral Large 3mistral/mistral-large-3Automanual only262.1K$0.50 / $1.50
- MiniMax M3minimax/minimax-m3Autoconversation · standardwriting · lightcoding · standardanalysis · heavyextraction · heavy512K$0.30 / $1.20
- Llama 4 Scoutmeta/llama-4-scoutAutomanual only128K$0.17 / $0.66
- Llama 4 Maverickmeta/llama-4-maverickAutomanual only128K$0.24 / $0.97
- GLM 5.3 FlashXzai/glm-5.3-flashxAutomanual only1M$0.37 / $1.25
- GLM 5.3 Flashzai/glm-5.3-flashAutoconversation · standardcoding · heavyanalysis · heavyextraction · standard1M$0.15 / $0.50
- GLM 5.3zai/glm-5.3Autoconversation · standardwriting · heavycoding · heavyanalysis · heavyextraction · standard1M$1.40 / $4.40
Image models
- GPT Image 2Defaultopenai/gpt-image-2square · landscape · portrait
- GPT Image 1.5openai/gpt-image-1.5square · landscape · portrait
- GPT Image 1 Miniopenai/gpt-image-1-minisquare · landscape · portrait
- FLUX.2 Maxbfl/flux-2-maxsquare · landscape · portrait
- FLUX.2 Probfl/flux-2-prosquare · landscape · portrait
- Grok Imagine Image 2.0spacexai/grok-imagine-image-2.0square · landscape · portrait
Embedding
voyage/voyage-41024 dimensions