gpt-5.6-luna

OpenAI's July 30, 2026 product post explicitly names GPT‑5.6 Luna as "our fastest and most affordable model" in the GPT‑5.6 family and announces an 80% price cut, positioning Luna for high-volume enterprise work that can use tools and complete multi-step workflows. The post frames Luna's value as frontier-level intelli Because this is a first-party OpenAI announcement naming the exact "GPT‑5.6 Luna" version, it provides the strongest version-exact evidence among the candidates. The excerpt does not discuss SAP AI Core hosting specifically, so any inference that Luna is "on" SAP AI Core must come from the subject metadata, not from th

CodeRabbit's July 9, 2026 engineering analysis of GPT-5.6 for coding agents and code review identifies Luna as "the fastest and lowest-cost tier" in the GPT-5.6 family, alongside Sol (flagship) and Terra (balanced), and frames the release as enabling work to be routed by depth rather than treating GPT-5.6 as a single m The blog is a vendor publication with a commercial framing (CodeRabbit cites a $143M funding round in the excerpt) and provides no Luna-specific benchmark results, evaluation methodology, or SAP AI Core hosting/API detail. Its contribution is limited to confirming Luna's creator-attributed positioning as the lowest-rea

The llm-stats aggregator page for GPT-5.6 Luna (published July 9, 2026) provides the most Luna-specific quantitative evidence in the candidate set: an overall LLM Stats Score of 45.1 at a blended price of $0.25 per million tokens, an overall rank of 32, top-10% standing in coding (21 of 266), and below-top-half standin Per-benchmark detail in the excerpt covers Connectors (OpenAI internal production tool-use pass rate) and HealthBench Consensus, among others. The page is third-party rather than creator-published, does not identify which inference hosts contributed to scores, and contains no SAP AI Core-specific gateway, API, or prici

On July 9, 2026, OpenAI launched the GPT-5.6 family to general availability following the June limited preview, with Sol as the flagship, Terra as the balanced tier, and Luna as the most cost-efficient model. The excerpt explicitly names Luna, confirms GA across ChatGPT, Codex, and the OpenAI API, and includes an updat For Luna specifically, this first-party post confirms general availability, identifies it as the cost-efficient tier of the GPT-5.6 family, and records an 80% Luna price reduction on July 30, 2026, though the excerpt does not state the new Luna rate or Luna-specific benchmark numbers. The source is OpenAI's own GA anno

VentureBeat's June 26, 2026 coverage of OpenAI's preview describes Luna as the GPT-5.6 variant aimed at "faster, lower-cost everyday work like summarization, drafting and routine automation," while noting it "performs near GPT-5.5 levels on several tests despite being positioned as the fastest and lowest-cost model." T The article's primary focus is policy and access context rather than Luna-specific technical assessment, and it contains no SAP AI Core integration, regional availability, or API detail. Its contribution is third-party corroboration of Luna's family placement, capability positioning, and the preview-stage access condit

OpenAI's June 26, 2026 preview post introduces the GPT‑5.6 series and explicitly enumerates the three tiers: "Sol, our flagship model; Terra, a balanced model for everyday work; and Luna, a fast and affordable model." It states that Luna "brings strong capability at our lowest cost" and that GPT‑5.6 Sol launches with t The post centers on Sol's preview and safety posture, so Luna-specific capability detail is light, but the version-exact naming of "Luna" as a tier in the GPT‑5.6 family is directly supported by the page excerpt. Like the price-performance post, it does not address SAP AI Core hosting, so provider-level claims about wh

OpenRouter's listing for the GPT-5.6 Luna batch variant reports a release date of July 9, 2026, a February 2026 knowledge cutoff, a 1M-token context window, and pricing of $0.10 per 1M input tokens, $0.60 per 1M output tokens, with a $0.01 per 1M cache-read price. It frames Luna as a fast, cost-efficient model for high The page breaks down per-provider benchmark scores on GPQA Diamond and TAU-Bench across Azure (multiple regions), OpenAI direct, Amazon Bedrock, and OpenRouter's auto-routing tier, giving developers a cross-provider reference point for Luna quality. It also reports 0.00% uptime and 0.00% availability over the trailing