Models › Text › GPT-6 Luna Pro
OpenAI · Model details, evidence, and API facts
GPT-6 Luna Pro runs GPT-6 Luna in pro reasoning mode
Starting provider API rate; context tiers apply
See what GPT-6 Luna Pro does best, then review its limits, benchmarks, and pricing below
GPT-6 Luna Pro tackles harder everyday work with deeper reasoning
Every GPT-6 Luna Pro result below comes from a named source, labeled clearly by protocol, so each score reflects how it was measured
| Benchmark | Score | Evaluation details | Source |
|---|---|---|---|
AIM Text-to-SQL Measures GPT-6 Luna Pro on 759 text-to-SQL questions scored by exact results. |
44.2% SQL accuracy |
Version: AIM Text-to-SQL • Split: 759 questions, one run per model • Scored by exact result matching after reference review • Slightly below GPT-6 Luna at 45.4% — pro mode does not guarantee higher scores | Methodology AIMultiple Independently run |
DGF-Bench Tests GPT-6 Luna Pro on 27 fixed attacks probing AI replacement decisions. |
85.2 DGF score (out of 100) |
Version: DGF-Bench • Split: 27 fixed attacks, 34 clean gates • Evaluated independently through OpenRouter, results of September 27, 2026 • Blocked 23 of 27 attacks, clean 34/34 • Ranked #4 of 6 frontier models | Methodology Frontier Labs DGF-Bench Independently run |
Official A Tests GPT-6 Luna Pro on 157 strict tasks spanning coding and reasoning. |
64.3% Accuracy (thinking off) |
Version: Official A strict_v01 • Split: 157 tests • Run by SMF Clearinghouse on September 22, 2026, same day and harness as Claude Opus 5.5 (96.2%) • Reasoning effort: none • Cost $0.05, 8.4 minutes, zero errors | Methodology SMF Clearinghouse Independently run |
GPQA Diamond Tests GPT-6 Luna Pro on 198 Google-proof graduate-level questions in biology, chemistry, and physics to probe expert-level scientific reasoning. |
Not published Accuracy |
No published score for the Pro configuration. Version, split, evaluator, and reasoning effort are pending a published Pro-specific result. | No verified Pro-specific source available |
Terminal-Bench 4.0 Tests GPT-6 Luna Pro on complex terminal-based tasks spanning software engineering, system configuration, and data analysis, rewarding long-horizon tool use and error recovery. |
Not published Independent task success rate |
No published score for the Pro configuration. Version, split, harness, and reasoning effort are pending a published Pro-specific result. | No verified Pro-specific source available |
Arena rankings capture how people vote between anonymous responses. GPT-6 Luna Pro’s preference standing is reported per leaderboard, since results can vary widely by task
FILL — no LMArena rating found for GPT-6 Luna Pro in published leaderboards as of Nov 6, 2026. Do NOT copy GPT-6 Luna’s Arena results.
GPT-6 Luna Pro offers a 1,050,000-token context window, 128,000 output tokens, file input, and pro reasoning, with full API pricing and tiers shown below
| Model | |
| Provider | OpenAI |
| Release date | Sep 22, 2026 |
| Catalog added | [Pyxa catalog date] |
| Context window | 1,050,000 tokens |
| Maximum output | 128,000 tokens |
| Knowledge cutoff | 2026-05-18 |
| License | Proprietary |
| Capabilities and API | |
| Input types | Text, Image, File |
| Output types | Text |
| Reasoning | Supported (pro reasoning mode) |
| Provider endpoint tool parameters | Accepts tool parameters |
| Base provider API input | $0.10 / 1M tokens |
| Base provider cached input (read) | $0.01 / 1M tokens |
| Base provider API output | $0.50 / 1M tokens |
| Provider tier above 272,000 prompt tokens | Input $0.20 / 1M • Output $0.75 / 1M • Cached input $0.02 / 1M |
Provider token rates are context-dependent. The base rates apply below the listed prompt thresholds; the matching tier applies at or above each threshold. These are reference provider API rates, not Pyxa AI checkout charges — access in Pyxa AI follows the allowances of your Pyxa AI plan.
Each GPT-6 Luna Pro benchmark and specification links to its source, so you can check where every figure comes from
Discover where GPT-6 Luna Pro performs best, from harder everyday tasks to fast daily high-volume work at scale
Put GPT-6 Luna Pro to work on chat, classification, and agentic tasks