npub1kc…f0vwv on Nostr: OpenAI Unveils GPT-6 Sol & Luna, Challenging Anthropic's Claude 5.5 with Major Price ...
OpenAI Unveils GPT-6 Sol & Luna, Challenging Anthropic's Claude 5.5 with Major Price Cuts
OpenAI has launched two new AI models, GPT-6 Sol and GPT-6 Luna, directly after Anthropic's release of Claude Opus 5.5. These models offer high-performance reasoning at significantly reduced costs, aiming to compete with Anthropic's Claude AI ecosystem.
Sol and Luna are positioned as cost-effective alternatives to GPT-6 Astra. Sol competes with Opus, while Luna targets the Haiku model. OpenAI has reduced API costs by 50% compared to previous GPT-5.6 rates: GPT-6 Sol is $2/million input tokens and $10/million output tokens; GPT-6 Luna is $0.10/million input tokens and $0.50/million output tokens.
Benchmarking shows GPT-6 Sol excelling in cost-to-performance ratios, outperforming Claude Opus 5 on AutomationBench and Agents' Last Exam. A new "Reasoning Effort" toggle allows users to balance speed and cost against accuracy. These models are initially available for ChatGPT subscribers, with wider integration planned.
https://cryptovka.com/news/openai-debuts-gpt-6-sol-and-luna-models-to-rival-anthropic-s-claude-5-5Published at
2026-09-22 19:02:35 UTCEvent JSON
{
"id": "e2269ecd158f54e45765ae234ddcecfcb91d129bc78d8dcc14bad91ef4a1cd9e",
"pubkey": "b63581fed371e76c0e77ca94a868b56f191c3b06240c0205bf1e9098d2cccc49",
"created_at": 1790103755,
"kind": 1,
"tags": [],
"content": "OpenAI Unveils GPT-6 Sol \u0026 Luna, Challenging Anthropic's Claude 5.5 with Major Price Cuts\n\nOpenAI has launched two new AI models, GPT-6 Sol and GPT-6 Luna, directly after Anthropic's release of Claude Opus 5.5. These models offer high-performance reasoning at significantly reduced costs, aiming to compete with Anthropic's Claude AI ecosystem.\n\nSol and Luna are positioned as cost-effective alternatives to GPT-6 Astra. Sol competes with Opus, while Luna targets the Haiku model. OpenAI has reduced API costs by 50% compared to previous GPT-5.6 rates: GPT-6 Sol is $2/million input tokens and $10/million output tokens; GPT-6 Luna is $0.10/million input tokens and $0.50/million output tokens.\n\nBenchmarking shows GPT-6 Sol excelling in cost-to-performance ratios, outperforming Claude Opus 5 on AutomationBench and Agents' Last Exam. A new \"Reasoning Effort\" toggle allows users to balance speed and cost against accuracy. These models are initially available for ChatGPT subscribers, with wider integration planned.\n\nhttps://cryptovka.com/news/openai-debuts-gpt-6-sol-and-luna-models-to-rival-anthropic-s-claude-5-5",
"sig": "6c6bf5f97f1d5a3a5283f70b30f410ae77a569ef664858dc34a298c18f517484f9386624fbed0284d9a4f0d7b2959832f3f7be003d53b1ce5d48a701ea37280e"
}