GPT-6 Luna
openai/gpt-6-luna-20260922
OpenAI's GPT-6 Luna model accepts file, image, and text inputs and generates text outputs. It supports a context length of 1,050,000 tokens, tools, and reasoning capabilities. It is a proprietary model released on September 22, 2026.
- tool use
- structured outputs
- reasoning
At a glance
Specifications
What the API gives you
- Context window
- 1.1M tokens
- Max output
- 128K tokens
- Input price
- $0.100 / 1M tokens
- Output price
- $0.50 / 1M tokens
- Cache read
- $0.010 / 1M tokens
- Modalities
- file, image, text → text
- Knowledge cutoff
- —
- Released
- Sep 22, 2026
Measured evidence
Public benchmark record
- FACTS Grounding
- 76.7%2026-10-27evaluated 2026-10-27
- FrontierMath Tiers 1-3 (Epoch AI)‡
- 78.9% ± 2.42026-09-22max effort, evaluated 2026-09-22
- GPQA Diamond (Epoch AI)‡
- 90.5% ± 1.92026-09-22max effort, evaluated 2026-09-22
- LiveBench Agentic Coding
- 51.2%2026-10-18
- LiveBench Coding
- 79.0%2026-10-18
- LiveBench Data Analysis
- 73.4%2026-10-18
- LiveBench Instruction Following
- 55.9%2026-10-18
- LiveBench Language
- 73.8%2026-10-18
- LiveBench Math
- 89.1%2026-10-18
- LiveBench Reasoning
- 81.8%2026-10-18
- LMArena Agent‡
- 1.62026-11-11max effort
- LMArena Agent Bash Recovery‡
- 5.22026-10-08max effort
- LMArena Agent Steerability‡
- 3.52026-11-11max effort
- LMArena Agent Task Outcome‡
- -0.42026-11-11max effort
- LMArena Agent Tool Hallucination‡
- 12026-11-11max effort
- LMArena Creative Writing‡
- 1359 Elo2026-11-11max effort
- LMArena Instruction Following‡
- 1403 Elo2026-11-11max effort
- LMArena Long Query‡
- 1403 Elo2026-11-11max effort
- LMArena Math‡
- 1411 Elo2026-11-11max effort
- LMArena Search Factuality‡
- 1186 Elo2026-11-09max effort
- LMArena Text‡
- 1441 Elo2026-11-11max effort
- LMArena Text Factuality‡
- 1466 Elo2026-11-11max effort
- LMArena Vision‡
- 1235 Elo2026-11-09max effort
- LMArena Vision Diagrams‡
- 1266 Elo2026-11-09max effort
- LMArena Vision OCR‡
- 1248 Elo2026-11-09max effort
- LMArena WebDev‡
- 1577 Elo2026-11-11max effort
- LMArena WebDev Image-to-Web‡
- 1566 Elo2026-11-10max effort
- OTIS Mock AIME (Epoch AI)‡
- 98.9% ± 0.72026-09-22max effort, evaluated 2026-09-22
- SimpleQA Verified (Epoch AI)‡
- 41.4% ± 1.62026-09-22max effort, evaluated 2026-09-22
- Terminal-Bench (4.1)‡
- 15.2%2026-11-15Codex, max effort, evaluated 2026-11-03
LMArena data licensed CC-BY-4.0. Aider (Apache-2.0) and SWE-bench (MIT) are open. Design Arena uses blind human preference. ‡ Submitted scaffold plus model result. Compare the harness and effort settings, not only the model name.
Task evidence
How broad is the support?
Each status is exact to this model version and the selected task. Official release claims remain provisional.
- Best LLM for Competition MathEstablished evidence
100% of intended weighted task evidence across 4 independent sources. 6 of 6 configured evaluation feeds currently have results.
- Best LLM for Business WritingEstablished evidence
100% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for Logic & Problem SolvingEstablished evidence
100% of intended weighted task evidence across 3 independent sources. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Marketing CopyEstablished evidence
100% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for Chat & RoleplayEstablished evidence
100% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for Creative WritingEstablished evidence
100% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
Show all 45 tasks
- Best LLM for WritingEstablished evidence
100% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for Roleplay & CharacterEstablished evidence
100% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for Math & ReasoningEstablished evidence
100% of intended weighted task evidence across 5 independent sources. 7 of 7 configured evaluation feeds currently have results.
- Best LLM for a General AssistantEstablished evidence
100% of intended weighted task evidence across 3 independent sources. 6 of 6 configured evaluation feeds currently have results.
- Best LLM for Grounded RAGEstablished evidence
81% of intended weighted task evidence across 3 independent sources. 5 of 7 configured evaluation feeds currently have results.
- Best LLM for Code CompletionEstablished evidence
79% of intended weighted task evidence across 2 independent sources. 3 of 3 configured evaluation feeds currently have results.
- Best LLM for Autonomous AgentsEstablished evidence
67% of intended weighted task evidence across 3 independent sources. 6 of 6 configured evaluation feeds currently have results.
- Best LLM for Document SummarizationEstablished evidence
65% of intended weighted task evidence across 3 independent sources. 5 of 6 configured evaluation feeds currently have results.
- Best LLM for SummarizationEstablished evidence
65% of intended weighted task evidence across 3 independent sources. 5 of 6 configured evaluation feeds currently have results.
- Best Open-Weight LLM for Self-HostingEstablished evidence
64% of intended weighted task evidence across 2 independent sources. 2 of 3 configured evaluation feeds currently have results.
- Best LLM for Image UnderstandingEstablished evidence
61% of intended weighted task evidence across 2 independent sources. 2 of 3 configured evaluation feeds currently have results.
- Best LLM for Agentic CodingEstablished evidence
60% of intended weighted task evidence across 3 independent sources. 7 of 7 configured evaluation feeds currently have results.
- Best LLM for Fast SummariesEstablished evidence
58% of intended weighted task evidence across 2 independent sources. 2 of 3 configured evaluation feeds currently have results.
- Best Value LLMEstablished evidence
58% of intended weighted task evidence across 2 independent sources. 2 of 3 configured evaluation feeds currently have results.
- Best LLM for RAG & Long ContextEstablished evidence
56% of intended weighted task evidence across 2 independent sources. 5 of 6 configured evaluation feeds currently have results.
- Best LLM for AI AgentsEstablished evidence
50% of intended weighted task evidence across 3 independent sources. 9 of 9 configured evaluation feeds currently have results.
- Best LLM for Literary TranslationEstablished evidence
50% of intended weighted task evidence across 2 independent sources. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for Everyday TranslationEstablished evidence
50% of intended weighted task evidence across 2 independent sources. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for TranslationEstablished evidence
50% of intended weighted task evidence across 2 independent sources. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for CodingEstablished evidence
49% of intended weighted task evidence across 3 independent sources. 7 of 7 configured evaluation feeds currently have results.
- Best LLM for Hermes AgentEstablished evidence
48% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for OpenHandsEstablished evidence
48% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for GooseEstablished evidence
48% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for OpenClawEstablished evidence
48% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best Local LLM for CodingEstablished evidence
47% of intended weighted task evidence across 2 independent sources. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for Frontend & UIEstablished evidence
45% of intended weighted task evidence across 2 independent sources. 6 of 6 configured evaluation feeds currently have results.
- Best Cheap LLM for High VolumeEstablished evidence
40% of intended weighted task evidence across 2 independent sources. 2 of 3 configured evaluation feeds currently have results.
- Cheapest Capable LLMEstablished evidence
40% of intended weighted task evidence across 2 independent sources. 2 of 3 configured evaluation feeds currently have results.
- Best LLM for JSON & Schema OutputEstablished evidence
35% of intended weighted task evidence across 2 independent sources. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for Structured Data ExtractionEstablished evidence
35% of intended weighted task evidence across 2 independent sources. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for Debugging & Code ReviewEstablished evidence
33% of intended weighted task evidence across 2 independent sources. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Tool & Function CallingProvisional evidence
46% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from Berkeley Function Calling or Structured-output eval. 6 of 7 configured evaluation feeds currently have results.
- Best LLM for OCR & DocumentsProvisional evidence
45% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 3 of 5 configured evaluation feeds currently have results.
- Best LLM for Charts & DiagramsProvisional evidence
45% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 3 of 5 configured evaluation feeds currently have results.
- Best LLM for Vision & DocumentsProvisional evidence
40% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 4 of 6 configured evaluation feeds currently have results.
- Best LLM for Document ParsingProvisional evidence
37% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 3 of 5 configured evaluation feeds currently have results.
- Best LLM for High-Volume ExtractionProvisional evidence
21% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from Route reliability or Structured-output eval. 1 of 3 configured evaluation feeds currently have results.
- Best LLM for Massive ContextProvisional evidence
20% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or LongBench v2. 2 of 3 configured evaluation feeds currently have results.
- Best LLM for Computer & Browser UseProvisional evidence
17% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from Berkeley Function Calling or OSWorld Verified. 3 of 3 configured evaluation feeds currently have results. Official release evidence is also tracked provisionally.
Official release evidence · provisional
What the launch material actually says
Verbatim claims from the model provider. These help on launch day but do not count as independent confirmation.
“On DeepSWE v1.1, which tests performance on complex software-engineering tasks in real codebases, GPT‑6 Sol at max effort scores 68.8%”
OpenAI · Introducing GPT-6 Sol and Luna | OpenAI“At high effort, GPT‑6 Luna improves on its predecessor by 5.4 percentage points at 58% lower cost per task.”
OpenAI · Introducing GPT-6 Sol and Luna | OpenAI“GPT‑6 Luna at max effort scores 66.6%, comparable to Claude Opus 5 and Fable 5 at medium effort.”
OpenAI · Introducing GPT-6 Sol and Luna | OpenAI“GPT‑6 Luna (max) is able to exceed GPT‑5.6 Sol (medium) at one tenth of its cost.”
OpenAI · Introducing GPT-6 Sol and Luna | OpenAI
Recommendation coverage
Where this model appears
Catalog data via OpenRouter and models.dev. Last checked within the past 6 hours.