GPT-6 Luna
gpt-6-lunaopenai-gpt-6-luna- 📏 1.05M-token context: 922K input, 128K output
- ⚡ OpenAI's most efficient GPT-6 tier for high-volume work
- 👁️ Accepts text and image input, returns text
- 🔧 Function calling, web search, Responses API tools
- 🧠 Selectable reasoning effort from none up to max
- 🆕 Successor to GPT-5.6 Luna, launched alongside GPT-6 Sol
- 💬 Aimed at chat, classification, lightweight agentic workflows
- 🎯 Served via Responses and Chat Completions APIs
OpenAI is an American artificial intelligence research organization headquartered in San Francisco, structured as both a for-profit public benefit corporation and a nonprofit foundation. The lab developed the GPT family of large language models, the DALL-E image generation…
Explore 29 more models by OpenAI →GPT-6 Luna is the smallest, most cost-efficient member of OpenAI's GPT-6 lineup, released in September 2026 alongside GPT-6 Sol and sitting below the flagship GPT-6 Astra. OpenAI's own model guidance positions it for cost-sensitive, high-volume workloads, while Sol balances intelligence and cost and Astra handles the most complex reasoning and coding. It accepts text and images, emits text, and exposes a 1.05 million-token context window with up to 128K output tokens.
Compared with its direct predecessor, GPT-5.6 Luna, the continuity is notable: both share the same 1.05M-token window and the same selectable reasoning-effort ladder running from none through low, medium, high and max, plus vision input, function calling and hosted tools such as web search. The GPT-6 release refreshes this efficiency tier within the newer generation, so the same integration surface — Responses and Chat Completions — carries over for existing 5.6 Luna deployments.
Practically, Luna suits classification, extraction, routing, retrieval over very long documents, and focused agent steps, with reasoning effort dialled up when a task needs more deliberation and dialled down for latency-sensitive, high-throughput traffic. Teams needing deeper multi-step reasoning or heavier coding work are pointed by OpenAI's documentation toward Sol or Astra instead, keeping Luna as the volume workhorse of the family.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Routing | Input $/M | Cached $/M | Output $/M | Categories | API |
|---|---|---|---|---|---|---|---|
| ▲ Apex Ant 0x73b4…e736 | 71.90 | #4 | $0.10 | $0.01 | $0.4688 | chat,reasoning,premium,frontier,vision,multimodal,long-context,research,coding,agents,tasks,tools,structured,writing,study,finance,legal | openai-chat-completions |
| D5V1N2 0xd5e7…7be0 | 67.45 | #5 | $0.11 | $0.011 | $0.55 | chat,coding,agent,reasoning,vision,large-context,openai,router,fallback | openai-chat-completions |
| GesundAI 0x983e…d826 | 66.98 | #2 | $0.03 | $0.003 | $0.15 | chat,code,reasoning,multimodal | openai-responses |
| TokenSale 0x2060…742d | 65.23 | #1 | $0.018 | $0.0018 | $0.09 | chat,coding,fast | openai-chat-completions |
| Super Seeder 0xd19f…41f3 | 63.77 | #3 | $0.0619 | $0.0062 | $0.3094 | chat,reasoning,vision,multimodal,tools,long-context,cheap | openai-chat-completions |
| ZLKPro-Api 0x0b0b…f446 | 58.35 | gated | $0.01 | $0.01 | $0.03 | agent,chat,text,reasoning,research,smart,long-context,multimodal | openai-chat-completions |
| bartly.eth64.de 0x666e…4666 | 53.43 | gated | $0.0023 | $0.0002 | $0.0116 | chat,coding,reasoning,vision,multimodal,web-search,function-calling | openai-chat-completions |
| Fire Ant 🔥🐜 0xbe05…bc5d | 50.00 | gated | $0.1034 | $0.0093 | $0.5168 | chat,coding,reasoning,tools | — |
| Best 0x472f…69fd | 49.76 | gated | $0.0023 | $0.0002 | $0.0115 | chat,coding | openai-chat-completions |
| Hana Gateway ✅ 0x4ae1…117b | 45.41 | gated | $0.01 | $0.01 | $0.06 | chat,coding,reasoning,tools | openai-chat-completions |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the AntSeed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on AntSeed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.