BlogLLM Benchmark Scores as Published, Side by Side
GPT-5.6 Luna: Benchmark Scores as Published
Every result the catalog holds for GPT-5.6 Luna: 107 tests in 16 areas, from 8 pages. 91 printed by OpenAI itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer OpenAIReleased 2026-07-09Context window 1.05MMax output 128KInput Text, imageOutput TextWeights ClosedData version 2026103001, as of 2026-10-30
Knowledge Work
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| AutomationBench | 14.9% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
| GDPval-AA v2 | 1591.8 | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
| APEX-Agents | 28.6% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| GDPval-AA (Elo) | 1569 | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| Management Consulting Tasks (Internal) | 35.4% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| Big Finance Bench | 36% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-Bench Pro | 62.7% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 2 more pages | - |
| Terminal-Bench 2.1 | 84.7% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 2 more pages | - |
| DeepSWE v1.1 | 67.2% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
| Artificial Analysis Coding Agent Index v1.1 | 74.6 | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
ML Engineering
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Internal Research Debugging Evaluation | 50.8% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| MLE-bench | 47.6% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15 | - |
| KernelGen 1P | 22.4% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| NanoGPT | 1.66% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| PostTrainBench Lite | 29.6% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| RSI Index (OpenAI) | 41.9% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 92.3% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
| SciCode | 52.5% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| CritPt | 21% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| FrontierMath Tier 4 (v2) | 58.5% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| FrontierMath Tier 1-3 (v2) | 78.6% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| GeneBench Pro | 10.8% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| LifeSciBench | 51.2% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| MedChemBench (Internal) | 30.4% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| AA-LCR | 78.3% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| MRCR v2 8-needle 128K | 74.8% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15 | Context: 128k (average) |
| GraphWalks BFS 1M (F1) | 51.2% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Context: 1M |
| GraphWalks BFS 256k (F1) | 81.3% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Context: 256k |
| MRCR v2 8-needle 512K-1M | 41.3% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Context: 512K-1M |
| MRCR v2 8-needle 256K-512K | 41.3% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Context: 256K-512K |
Computer Use
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OSWorld-Verified | 72.6% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15 | - |
| OSWorld 2.0 | 45.6% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
Multimodal Understanding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| CharXiv Reasoning (no tools) | 82.7% | Google, Gemini 3.6 Flash model cardchecked 2026-10-15 | Tools: none |
| MMMU-Pro | 78.4% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Tools: none |
| GDP.pdf | 22.7% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| MMMU-Pro (with tools) | 79.5% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Tools: with tools |
Cybersecurity
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| ExploitBench | 33.2% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Harness: ExploitBench API harness, 5 seeds, reasoning continuity (footnote 2) |
| ExploitGym | 12.4% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| SEC-Bench Pro | 48.9% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| Production Chat | 0.986 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Synthetic Agentic Envs | 1 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Capture the Flag challenges | 85.2% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 39.5% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max; Tools: none |
| AA-Omniscience Accuracy | 43% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| AA-Omniscience Non-Hallucination | 7% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| ARC-AGI-3 | 0.18% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Toolathlon | 53.4% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| τ³-bench Banking | 31.1% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| Toolathlon Verified | 67.5% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| MCPMark | 66.9% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
| Agents' Last Exam (score) | 50.3% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| WildClawBench | 50.4% | IFM, IFM K2-Horizon-375B-A23B model cardchecked 2026-10-15 | Effort: max |
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 83.3% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15; printed on 1 more page | - |
Healthcare
Biosecurity
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Dual Use Safe | 0.926 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Severe Safe | 0.946 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Benign Not overrefuse | 0.989 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 1 more page | - |
Safety and Alignment
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Age-restricted goods, services, and dangerous challenges / activities | 0.76 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Eating Disorders | 0.702 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Emotional reliance | 0.957 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Emotional Reliance | 0.931 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| extremism | 0.966 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Extremism | 0.981 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Gorefraction | 0.585 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| harms-erotic | 0.986 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| hate | 0.996 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Hate | 1 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Mental health | 0.989 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Non-Violent Illicit behavior | 0.993 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Self Harm | 0.99 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| self-harm | 0.99 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Self-harm | 0.905 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Self-harm (standard) | 0.954 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Sexual | 0.944 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Sexual Content | 0.922 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Sexual/minors | 0.974 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Violent Illicit behavior | 0.94 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 2 more pages | - |
| Gorefraction | 0.819 | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 1 more page | - |
| Respecting Warnings: unwanted persistencelower is better | 27% | OpenAI, GPT-6 Sol and GPT-6 Luna: October 2026 updatechecked 2026-10-19 | - |
| Circumventing restrictions (lower is better)lower is better | 0% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: With confirmation policy (default for users) |
| 0.4% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: Base Model (without confirmation policy) | |
| Data exfiltration (lower is better)lower is better | 8.5% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: With confirmation policy (default for users) |
| 15.4% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: Base Model (without confirmation policy) | |
| Destructive action (lower is better)lower is better | 2.4% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: Base Model (without confirmation policy) |
| 0% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: With confirmation policy (default for users) | |
| Overall misaligned outcome rate (lower is better)lower is better | 19.7% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: Base Model (without confirmation policy) |
| 7.2% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: With confirmation policy (default for users) | |
| Security weakening (lower is better)lower is better | 2.1% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: With confirmation policy (default for users) |
| 7.3% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: Base Model (without confirmation policy) | |
| Service disruption (lower is better)lower is better | 0.6% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | - |
| Unauthorized external communication (lower is better)lower is better | 7.3% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: With confirmation policy (default for users) |
| 15.6% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: Base Model (without confirmation policy) | |
| Unauthorized transactions (lower is better)lower is better | 33.5% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: Base Model (without confirmation policy) |
| 5.3% | OpenAI, GPT-6 Astra System Cardchecked 2026-10-25 | Configuration: With confirmation policy (default for users) | |
| Auto-review: exploitationlower is better | 0.3% | OpenAI, GPT-6.1 Sol system card addendumchecked 2026-10-19; printed on 1 more page | - |
Engineering and Design
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BenchCAD | 63.1% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
| BenchCAD (python tool) | 73.9% | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | Tools: python tool |
Preference and Composite Rankings
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Artificial Analysis Intelligence Index v4.1 | 51.2 | OpenAI, GPT-5.6: Frontier intelligence that scales with your ambitionchecked 2026-10-15 | - |
Price
API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked. Cost per task is measured by Vals AI.
| Price | Value | Printed by | Setup as printed |
|---|---|---|---|
| Input | $0.2 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.4 per 1M tokens above 272,000 tokens |
| Cached input | $0.02 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.04 per 1M tokens above 272,000 tokens |
| Cache write | $0.25 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.5 per 1M tokens above 272,000 tokens |
| Output | $1.2 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $1.8 per 1M tokens above 272,000 tokens |
| Input, batch | $0.1 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.2 per 1M tokens above 272,000 tokens |
| Cached input, batch | $0.01 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.02 per 1M tokens above 272,000 tokens |
| Cache write, batch | $0.125 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.25 per 1M tokens above 272,000 tokens |
| Output, batch | $0.6 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.9 per 1M tokens above 272,000 tokens |
| Input, flex | $0.1 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.2 per 1M tokens above 272,000 tokens |
| Cached input, flex | $0.01 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.02 per 1M tokens above 272,000 tokens |
| Cache write, flex | $0.125 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.25 per 1M tokens above 272,000 tokens |
| Output, flex | $0.6 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.9 per 1M tokens above 272,000 tokens |
| Input, fast | $0.4 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.8 per 1M tokens above 272,000 tokens |
| Cached input, fast | $0.04 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $0.08 per 1M tokens above 272,000 tokens |
| Cache write, fast | $0.5 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $1 per 1M tokens above 272,000 tokens |
| Output, fast | $2.4 | OpenAI, OpenAI API pricingchecked 2026-10-15 | Context: up to 272,000 tokens; Long context: $3.6 per 1M tokens above 272,000 tokens |
| Cost per taskVals AI, USD, lower is better | $0.73 | Vals AI, Vibe Code Bench v1.1 leaderboardchecked 2026-10-27 | Harness: OpenHands |
More from OpenAI
- GPT-6.1 Terra
- GPT-6.1 Sol
- GPT-6 Luna
- GPT-6 Sol
- GPT-6 Astra
- GPT-5.6 Sol
- GPT-5.6 Terra
- GPT-5.5
- GPT-5.5 Pro
- GPT-5.4 mini
- GPT-5.4 nano
- GPT-5.4
- GPT-5.4 Pro
- GPT-5.3-Codex
- GPT-5.2-Codex
- GPT-5.2
- GPT-5.2 Instant
- GPT-5.2 Pro
- GPT-5.1-Codex-Max
- GPT-5.1
- GPT-5.1 Instant
- GPT-5-Codex
- GPT-5
- GPT-5 mini
- GPT-5 Pro
- gpt-oss-120b
- gpt-oss-20b
- ChatGPT agent
- codex-1
- o3
- OpenAI o4-mini
- GPT-4.1
- GPT-4.5
- Deep research
- o3-mini
- o1
- o1-mini
- GPT-4o
- GPT-5.5 Instant
- gpt-5.6-cyber
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.