Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
OpenAI · Proprietary · rank 49 · bench-align-v5
GPT-6 Luna is a general-purpose language model. It ranks in the top quarter for multimodal tasks (top 20% of 136 scored models), by BenchLM's published scores. It accepts text, images and PDFs and produces text.
GPT-6 Luna by OpenAI scores 61.3 overall, costs $0.10 per 1M input tokens and $0.50 per 1M output tokens, and carries a context window of 1.05M tokens — observed 2026-11-21.
Seven axes from the ranking source. A missing axis is drawn as a gap.
Every published price component, including cache reads and writes.
Lifecycle events the source associates with this model, plus what this profile does not claim.
Each axis shows the ranking source’s score and, beside it, BenchLM’s separate verified score for that axis. The two are published side by side and are not interchangeable.
Measured values with the date they were observed. Nothing is inferred from a sibling model or a provider claim.
Time to first token: latency to the first answer chunk (Artificial Analysis, via BenchLM). Reasoning models include thinking time, so values can run to tens or hundreds of seconds.
Evidence key: ObservedUnavailable
| Measurement | Value | Observed | Last good | Evidence |
|---|---|---|---|---|
| Time to first token | Unavailable | Unavailable | Unavailable | |
| Throughput | 66 tok/s | 2026-11-21 | 2026-11-21 |
Regional or per-endpoint measurements appear only when the API supplies them; none are modelled here.
| Batch output / 1M tokens (OpenRouter batch) (base rate) | $0.25 |
| Batch cache read / 1M tokens (OpenRouter batch) (base rate) | $0.005 |
| Batch cache write / 1M tokens (OpenRouter batch) (base rate) | $0.0625 |
Batch API rate (asynchronous jobs), from OpenRouter's batch listing.
The standard input, output and cache rates above are the base rate. models.dev publishes higher rates for requests whose prompt exceeds the thresholds below. The batch rates have their own long-context tiers, from OpenRouter's batch listing, shown separately. The cost simulator prices a standard-rate request whose whole prompt is above a threshold entirely at the highest of these tiers it exceeds, and a batch request only at a batch tier; it never applies these tiers to batch rates. The cost example on this page uses the base rate: its 1,200-token prompt is below every threshold.
The batch rates above are the base batch rate. OpenRouter's batch listing, the same listing as those rates, publishes higher batch rates for requests whose total prompt is above the thresholds below. The cost simulator prices a batch request whose whole prompt is above a threshold entirely at the highest batch tier it exceeds.
10 conversations per day × 8 messages × 22 active days, 1200 input and 400 output tokens per message, no cache. This uses the same calculator as the cost simulator, so an unavailable applicable rate makes the example unavailable too.
Plans from the model's maker, and the API providers that list this exact model today. Links leave ModelScale; each provider sets its own prices and limits.
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| HealthBench (raw) HealthBench raw score | 50.0 | #7 | 5,000 multi-turn patient conversations | Realistic healthcare conversations | HealthBench: Evaluating Large Language Models Towards Improved Human Health |
| HealthBench (length-adjusted) HealthBench length-adjusted score | 54.5 | #7 | 5,000 multi-turn patient conversations | Realistic healthcare conversations | HealthBench: Evaluating Large Language Models Towards Improved Human Health |
| HealthBench Professional (raw) HealthBench Professional raw score | 61.2 | #9 | 525 physician-authored conversations | Professional clinical tasks | HealthBench Professional: Evaluating Large Language Models on Real Clinician Chats |
| Artificial Analysis Intelligence Index Artificial Analysis Intelligence Index | 27.9 | #71 | Cross-benchmark intelligence index | Display-only external reference | Artificial Analysis |
| AA-HLE Artificial Analysis Humanity's Last Exam | 38.5 | #57 | Expert-level questions | Frontier expert reasoning | Artificial Analysis Humanity's Last Exam Benchmark Leaderboard |
| AA-Omniscience Index Artificial Analysis Omniscience Index | -8.3 | #100 | Knowledge questions | Broad factual knowledge | AA-Omniscience: Knowledge and Hallucination Benchmark |
| AA-Omniscience Accuracy Artificial Analysis Omniscience Accuracy | 39.6 | #62 | Knowledge questions | Broad knowledge | Artificial Analysis model benchmarks |
| AA-Omniscience Hallucination Rate Artificial Analysis Omniscience Hallucination Rate | 79.3 | #88 | Knowledge questions | Factuality | Artificial Analysis model benchmarks |
| HealthBench Hard HealthBench Hard | 31.4 | #8 | 1,000 health prompts | Advanced health reasoning | Muse Spark Eval Methodology |
| HealthBench Professional HealthBench Professional | 60.8 | #10 | Clinician chat tasks | Professional clinical workflows | HealthBench Professional: Evaluating Large Language Models on Real Clinician Chats |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| Bug Hunt Bench Bug Hunt Bench | 18.3 | #26 | 105 planted bugs across two production TypeScript repositories | Blind production-repository bug finding and repair | Bug Hunt Bench method, caveats, and definitions |
| AA-SciCode Artificial Analysis SciCode | 54.6 | #32 | Scientific coding subproblems | Scientific programming | Artificial Analysis SciCode Benchmark Leaderboard |
| DeepSWE DeepSWE | 66.6 |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| FrontierMath v2 (Tiers 1-3) FrontierMath v2 Tiers 1-3 | 78.947 | #21 | 285 private advanced mathematics problems | From olympiad-plus to early research mathematics | FrontierMath v2 benchmark hub |
| FrontierMath v2 (Tier 4) FrontierMath v2 Tier 4 | 56.098 | #21 | 41 private extreme-difficulty mathematics problems | Research-level mathematics requiring hours or days of expert work | FrontierMath Tier 4 v2 leaderboard |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| ARC-AGI-1 ARC-AGI-1 Semi-Private Evaluation | 86.67 | #33 | Semi-private ARC-AGI-1 evaluation set | Abstract visual reasoning | ARC Prize leaderboard |
| ARC-AGI-2 Abstraction and Reasoning Corpus for AGI v2 | 59.305556 | #37 | Visual pattern completion and abstract reasoning | Expert-level — hardest public reasoning benchmark | ARC-AGI-2: A Harder General Intelligence Benchmark |
| ARC-AGI-3 Abstraction and Reasoning Corpus for AGI v3 | 0.10361 |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| AA Briefcase Artificial Analysis Briefcase | 1336 | #39 | Professional knowledge-work tasks | Professional work | Artificial Analysis Briefcase Benchmark Leaderboard |
| AA AutomationBench Artificial Analysis AutomationBench | 53.2 | #34 | Business-process automation tasks | Agentic automation | Artificial Analysis AutomationBench Benchmark Leaderboard |
| GDPval-AA GDPval-AA | 1443 |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| AA-MMMU-Pro Artificial Analysis MMMU-Pro | 79.7 | #25 | Multimodal academic reasoning | Frontier multimodal | Artificial Analysis MMMU-Pro Benchmark Leaderboard |
| ParseBench-1 ParseBench-1: document parsing for AI agents (LlamaIndex leaderboard, first release) | 64.13 | #23 | 2,078 human-verified enterprise pages from insurance, finance and government documents | No method is strong on all five dimensions; charts and visual grounding separate the field | ParseBench: A Document Parsing Benchmark for AI Agents |
| ExtractBench (LlamaIndex) |
| Benchmark | Value | Rank | Tasks | Difficulty | Provenance |
|---|---|---|---|---|---|
| SEC-Bench Pro SEC-Bench Pro | 34.2 | #15 | Security engineering tasks | Advanced cybersecurity | Introducing GPT-5.6 |
| MaxxBench MaxxBench (mutated DeepSWE tasks) | 51.7 | #11 | 89 mutated DeepSWE feature-implementation tasks, two attempts per model | Long-horizon software engineering under changed requirements | Why MaxxBench (unLAP research post) |
| SpireClimb SpireClimb (Slay the Spire 1 and 2) | 0.0 |
A second, independent score source. Epoch’s own figures in Epoch’s units, one line per model version Epoch evaluated (effort setting as Epoch labels it). They are not part of this site’s capability scores or rank.
| Benchmark | Variant | Score | Epoch date | Evidence |
|---|---|---|---|---|
| GPQA Diamond Accuracy on graduate-level science multiple-choice questions (best score across scorers). Unit: fraction (0–1); higher is better. | max | 0.9047 ± 0.0185 | 2026-09-22 | |
| FrontierMath Tiers 1-3 (v2) Share of private research-level math problems solved, tiers 1-3. Unit: fraction (0–1); higher is better. | max | 0.7895 ± 0.0242 | 2026-09-22 | |
| FrontierMath Tier 4 (v2) Share of private research-level math problems solved, tier 4. Unit: fraction (0–1); higher is better. | max | 0.561 ± 0.0785 | 2026-09-22 | |
| SimpleQA Verified Short factual questions answered correctly. Unit: fraction (0–1); higher is better. | max | 0.414 ± 0.0156 | 2026-09-22 |
Source: Epoch AI, CC BY 4.0 (licence), benchmark data read once a day.
People's votes between two anonymous answers on arena.ai, one board per task. LMArena's own ratings and ranks, one line per model version it rates; not part of this site's capability scores or rank.
| Board | Variant | Rating | Votes | Rank | Published | Evidence |
|---|---|---|---|---|---|---|
| Vision Arena People's votes between two answers about the same image, style-controlled. | Max | 1,234 ± 8 | 7,728 | 71 | 2026-11-19 |
Source: LMArena, CC BY 4.0 (licence), leaderboard data read once a day.
| #36 |
| 113 software engineering tasks across 91 repositories and 5 languages |
| Long-horizon software engineering |
| DeepSWE benchmark blog |
| VulcanBench Frontier v4 VulcanBench Frontier v4 | 71.9 | #17 | 23 behavioural-reconstruction coding tasks | Professional multi-file software engineering | VulcanBench Frontier v4 leaderboard |
| FrontierCode 2.0 (AA) FrontierCode 2.0, Artificial Analysis evaluation | 25.2 | #3 | 132 private software-engineering tasks | Frontier coding-agent quality | Artificial Analysis Intelligence Index |
| FrontierCode 2.0 FrontierCode 2.0 | 22.7 | #23 | 132 private software-engineering tasks | Frontier coding-agent quality | FrontierCode 2.0 |
| #26 |
| Interactive game-like tasks with hidden rules |
| Frontier agentic reasoning |
| ARC-AGI-3: A New Challenge for Frontier Agentic Intelligence |
| AA-LCR Artificial Analysis Long Context Reasoning | 83.3 | #13 | Long-context reasoning tasks | Long-context reasoning | Artificial Analysis model benchmarks |
| MLCR-AA Medical Long Context Reasoning (MLCR-AA) | 26.1 | #18 | Long, fragmented medical-record reasoning | Long-context medical reasoning | Medical Long Context Reasoning (MLCR-AA) |
| CritPt Critical Physics Tasks | 19.4 | #31 | Research-level physics questions | Research-level physics reasoning | CritPt Benchmark Leaderboard |
| ARC-AGI-3 Provider Adapter ARC-AGI-3 Semi-Private: Provider Adapter harness | 0.59 | #13 | — | — | ARC Prize verified ARC-AGI-3 results |
| BioBasicsBench BioBasicsBench: everyday wet-lab science | 37.6 | #4 | 100 routine wet-lab analysis tasks (88 multimodal) across five domains | Routine wet-lab data interpretation | Measuring how AI agents handle everyday wet-lab science |
| #42 |
| Agentic real-world work tasks |
| Professional agentic workflows |
| DeepSeek-V4 Technical Report |
| GDPval-AA GDPval-AA normalized | 46.6 | #37 | Economically valuable tasks | Professional agentic workflows | Artificial Analysis model benchmarks |
| AA Terminal-Bench 4.0 Artificial Analysis Terminal-Bench v4.0 | 12.6 | #38 | Terminal-based agent tasks | Agentic software engineering | Artificial Analysis Terminal-Bench v4.0 Benchmark Leaderboard |
| GDP.pdf Artificial Analysis GDP.pdf | 22.8 | #24 | Professional document-production tasks | Professional knowledge work | GDP.pdf Benchmark Leaderboard |
| AA-AnalystAgent Artificial Analysis AnalystAgent | 43.8 | #22 | Spreadsheet and document analysis questions | Business and data analysis | AA-AnalystAgent Benchmark Leaderboard |
| ExploitGym ExploitGym | 11.6 | #20 | 898 exploitation tasks | Advanced cybersecurity exploitation | ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks? |
| Terminal-Bench 2.1 (Vals) Terminal-Bench 2.1, Vals AI run | 73.0 | #22 | Difficult terminal tasks | Frontier agentic | Vals AI Terminal-Bench 2.1, Vals AI run leaderboard |
| Hermes Index Hermes Index: agentic capability and cost per task inside Hermes Agent | 33.9 | #15 | Four suites: Hermes Bench (150 tasks), TerminalBench 4 without its 4 GPU tasks, TerminalBench Science and SkillsBench | Frontier agentic | Hermes Index: an agentic model leaderboard |
| Hermes Bench Hermes Bench: 150 agent tasks graded on the files and state the agent leaves behind | 70.5 | #13 | 150 tasks across 25 categories | Frontier agentic | Hermes Bench (Hermes Index) |
| MazeBench MazeBench: visual spatial reasoning in a 3D open world | 0 | #28 | 100 hidden gems across 256 rooms in one open world | Long-horizon spatial planning, with single puzzles that can take over one hundred moves | MazeBench Results |
| Terminal-Bench 4.1 Terminal-Bench 4.1 | 15.20 | #36 | 66 professional computer-work tasks | Frontier autonomous knowledge work | Terminal-Bench 4.1 |
| AA AutomationBench (strict) Artificial Analysis AutomationBench, strict pass@1 | 20.2 | #2 | End-to-end business-workflow automation tasks | Agentic automation | Artificial Analysis AutomationBench Benchmark Leaderboard |
| OSWorld 2.1 Offline OSWorld 2.1 offline subset (82 tasks) | 48.9 | #10 | 82 offline computer-use workflows | Long-horizon professional workflows | OSWorld 2.1 release (xlang-ai/OSWorld-V2, tag osworld-v2.1) |
| AA Cyber Index Artificial Analysis Cyber Index | 52.7 | #4 | Three defensive cybersecurity evaluations | Real-world vulnerability discovery and patching | Artificial Analysis Cyber Index |
| CWE-Bench-AA Artificial Analysis CWE-Bench-AA | 56.7 | #11 | Real-world vulnerability patching tasks | Automated vulnerability remediation | CWE-Bench-AA Benchmark Leaderboard |
| DeepsecBench-AA Artificial Analysis DeepsecBench-AA | 23.5 | #13 | Vulnerability discovery in real codebases | Security code review | DeepsecBench-AA Benchmark Leaderboard |
| CyberGym-E2E-AA Artificial Analysis CyberGym-E2E-AA | 77.9 | #4 | 131 end-to-end vulnerability reproduction and patching tasks | Real-world cybersecurity | CyberGym-E2E-AA Benchmark Leaderboard |
| MKA Multimodal Knowledge Agent Benchmark | 36.8 | #17 | 200 multimodal professional tasks across 10 industry verticals | Expert multimodal knowledge work | MKA Benchmark Leaderboard |
| Agents' Last Exam (Linux) Agents' Last Exam, Linux subset | 57.1 | #2 | Linux agent tasks (subset size not published) | Advanced agentic work | Introducing Muse Ember |
| 91.01 |
| #11 |
| 370 documents, 4,869 pages, 8 business domains, 67 document types, each with a JSON Schema |
| Short documents are near solved; long record lists and grounding separate systems |
| ExtractBench: A Benchmark for Schema-Guided Enterprise Document Extraction |
| ExtractBench Pro ExtractBench Pro: long-tail document extraction challenges (LlamaIndex) | 0 | #9 | 100 synthetic documents, 2,622 pages, 86 document types, six challenge groups | The best system completes 58 of 100 documents perfectly; several complete none | ExtractBench Pro leaderboard and challenge descriptions |
| ParseBench-2 ParseBench-2: severity-aware document parsing (LlamaIndex leaderboard) | 74.46 | #12 | 4,317 page-level test cases from 2,645 documents in 13 datasets | Charts and visual grounding separate the field; several systems that lead on text and tables score near zero on charts | ParseBench-2: Severity-Aware Evaluation of Document Parsing |
| PerceptionBench (tools) PerceptionBench with tools | 62.8 | #3 | Fine-grained visual-perception tasks with tools | Fine-grained visual perception | Introducing Muse Ember |
| #11 |
| 60 Slay the Spire 1 runs (Ascension 20, Heart kill) and 75 Slay the Spire 2 runs (Ascension 10) per model |
| Maximum game difficulty, long-horizon computer use |
| SpireClimb (Epoch AI) |
| Chess Puzzles Chess Puzzles (Epoch AI) | 36.0 | #41 | 100 programmatically generated chess positions with a single best move | Novel chess positions, single best move | Chess Puzzles (Epoch AI) |
| Mystery Game Puzzles Mystery Game Puzzles (Epoch AI) | 7.0 | #70 | 100 single-best-move positions from an undisclosed game | Unfamiliar game positions, single best move | Mystery Game Puzzles (Epoch AI) |
| Furniture Assembly Furniture Assembly Benchmark (Epoch AI) | 44.2 | #19 | 60 photos of three IKEA builds, some with deliberate assembly mistakes | Spatial reasoning from technical diagrams to real photos | Furniture Assembly Benchmark (Epoch AI) |
| MirrorCode MirrorCode (Epoch AI and METR) | 26.7 | #33 | 30 reimplementation tasks (15 programs in two languages), three attempts each | Days-long software engineering without source code | MirrorCode (Epoch AI) |
| ExploitBench Internal Port ExploitBench — Internal Port (June–August 2026) | 0.0 | #9 | Recently disclosed V8 and Chrome vulnerabilities (June–August 2026) | Browser exploitation | GPT-6 Cyber System Card |
| ExploitGym Honeypot ExploitGym Honeypot out-of-scope exploitation rate | 0.00 | #5 | Hard ExploitGym tasks with an out-of-scope honeypot service | Cyber-agent scope adherence | GPT-6 Cyber: Advancing the frontier of aligned cybersecurity agents |
| FACTS Multimodal FACTS Multimodal | 38.7 | #21 | About 1,500 image questions: 711 public and 811 held out | Image-grounded factual answering | BenchLM |
| DAYJOB DAYJOB | 1.2 | #15 | 130 long-horizon tasks: 80 finance and 50 healthcare | Long-horizon professional work | BenchLM |
| DAYJOB Healthcare DAYJOB Healthcare | 1.2 | #10 | 50 long-horizon healthcare tasks across five subfields | Long-horizon clinical and healthcare operations work | BenchLM |
| Code2Video Code2Video | 1469 | #13 | Creative briefs rendered as motion-graphics video | Code-driven motion design | BenchLM |
| Chart Arena Chart Arena | 1400 | #14 | 13 chart briefs with small datasets, rated by human votes | Data visualisation | BenchLM |
| Joke Arena Joke Arena | 1481 | #12 | 100 joke prompts in 10 categories, rated by human votes | Humour writing | BenchLM |
| Pelican Arena Pelican Arena | 1534 | #15 | One prompt: an animated SVG of a pelican riding a bicycle, rated by human votes | Vector drawing and animation in code | BenchLM |
| ASCII Arena ASCII Arena | 1275 | #16 | ASCII drawings of subjects from Google's Quick, Draw! dataset, rated by human votes | Text-character drawing | BenchLM |
| Game Arena Go Game Arena Go | 474 | #32 | Head-to-head Go games on a 13 x 13 board | Strategic planning and rule-following | BenchLM |
| NYT Connections NYT Connections LLM Benchmark (extended version) | 68.7 | #56 | 940 NYT Connections puzzles, each with up to four added trick words | Wordplay, word knowledge and decoys that fit almost-but-not-quite | NYT Connections LLM Benchmark (Lech Mazur, GitHub) |
| StealthMark StealthMark Intelligence Benchmarks (overall score) | 39.1 | #31 | 8 original tasks: Newton's Mirror, Ground Truth, Latch, Foresight, Mona Lisa, Selfie, Beach and Raccoon | Chained physical reasoning, planning without tools and precise visual construction | StealthMark: The AI Benchmark Explained (Stealth Models) |