RankingGPT-5.6 Luna
GPT-5.6 Luna
GPT-5.6 Luna by OpenAI is a GPT-5.6 model. This profile collects the published benchmark results and reported effort settings for this model. Configurations are compared on the effort page. Ranks are on the leaderboard.
51 published benchmark measures · 176 measured effort results · 15 benchmark families contribute across 5 task areas. 5 capability estimates available in the full profile. See all results ↓ · Compare published benchmarks →
Model evidence summary
This profile combines published settings. It is not a runnable configuration or a leaderboard rank. Compare measured configurations →
Performance profile
Capabilities
Filled points are supported; hollow points are preliminary. Lines stop at unknown capabilities. Exact values and sources follow below.
Scores estimate outcomes against a shared reference panel; they are not accuracy percentages. Sparse or disconnected evidence cannot qualify an overall profile. Open a capability to inspect its evidence.
Reported effort · Not specified
Settings reported in this model's published benchmark results, including results outside the aggregate. Effort names are provider-specific. These are not API defaults or equal compute budgets.
- Not specified: 56 observations
Mixed settings means multiple settings occur in the evidence. Best across efforts means the source selected its best reported result across settings; it does not mean Max. Unspecified settings stay unknown. This model-summary chart combines reported settings. The leaderboard keeps identified configurations separate and excludes unknown effort.
Inspect each result and its source ↓ · Download effort evidenceScore contributions and missing evidence
15 contributing families across 5 capabilities. Fixed reference panels do not change when the catalog expands.
Capability is fitted jointly across families. Capability estimates below describe different task areas; their weighted sum is not the Capability score.
Results without reviewed compatibility or a reference match remain in the raw evidence below. Coverage counts only contributing results.
Model information & shareable badge
- Lab
- OpenAI
- Catalog status
- active
- Availability
- Public provider catalog; account and region restrictions may apply
- Family
- GPT-5.6
- Released
- —
- Context
- 1,050,000 tokens
- License
- proprietary
- Model card
- https://developers.openai.com/api/docs/models/all
- Default Capability family coverage
/badge/gpt-5.6-luna.svg
GPT-5.6 Luna measured results by effort setting
Each row is one published result for one effort setting. Evidence collected 2026-11-22. Compare the effort settings →
| Benchmark | Effort | Score | Source |
|---|---|---|---|
| AA-Briefcase Elo | High | 1169 | Artificial Analysis |
| AA-Briefcase Elo | Low | 740 | Artificial Analysis |
| AA-Briefcase Elo | Max | 1339 | Artificial Analysis |
| AA-Briefcase Elo | Medium | 933 | Artificial Analysis |
| AA-Briefcase Elo | Non-reasoning | 707 | Artificial Analysis |
| AA-Briefcase Elo | XHigh | 1267 | Artificial Analysis |
| AA-LCR v1.1 | High | 80.3% | Artificial Analysis |
| AA-LCR v1.1 | Low | 70.0% | Artificial Analysis |
| AA-LCR v1.1 | Max | 83.7% | Artificial Analysis |
| AA-LCR v1.1 | Medium | 75.0% | Artificial Analysis |
| AA-LCR v1.1 | Non-reasoning | 42.7% | Artificial Analysis |
| AA-LCR v1.1 | XHigh | 81.7% | Artificial Analysis |
| AA-Omniscience Accuracy | Max | 39.1% | Artificial Analysis |
| AA-Omniscience Index | Max | -17.1 | Artificial Analysis |
| Agents' Last Exam | XHigh | 30.3% | Agents' Last Exam |
| APEX-Agents-AA | Max | 35.8% | Artificial Analysis |
| ARC-AGI-2 | High | 29.3% | ARC Prize |
| ARC-AGI-2 | Low | 5.1% | ARC Prize |
| ARC-AGI-2 | Max | 59.5% | ARC Prize |
| ARC-AGI-2 | Medium | 7.4% | ARC Prize |
| ARC-AGI-2 | XHigh | 47.6% | ARC Prize |
| ARC-AGI-3 · standard | High | 0.1% | ARC Prize |
| ARC-AGI-3 · standard | Low | 0.2% | ARC Prize |
| ARC-AGI-3 · standard | Max | 0.2% | ARC Prize |
| ARC-AGI-3 · standard | Medium | 0.2% | ARC Prize |
| AutomationBench-AA | High | 35.6% | Artificial Analysis |
| AutomationBench-AA | Low | 11.7% | Artificial Analysis |
| AutomationBench-AA | Max | 50.2% | Artificial Analysis |
| AutomationBench-AA | Medium | 23.1% | Artificial Analysis |
| AutomationBench-AA | Non-reasoning | 5.0% | Artificial Analysis |
| AutomationBench-AA | XHigh | 42.6% | Artificial Analysis |
| Bug Hunt Bench | Max | 29.8% | Bug Hunt Bench |
| CritPt | High | 16.6% | Artificial Analysis |
| CritPt | Low | 2.6% | Artificial Analysis |
| CritPt | Max | 20.6% | Artificial Analysis |
| CritPt | Medium | 4.9% | Artificial Analysis |
| CritPt | Non-reasoning | 0.3% | Artificial Analysis |
| CritPt | XHigh | 20.6% | Artificial Analysis |
| DAYJOB Finance | Max | 1.8% | Surge AI |
| DAYJOB Healthcare | Max | 0.4% | Surge AI |
| DeepSWE v1.1 | High | 44.2% | Datacurve |
| DeepSWE v1.1 | Low | 1.5% | Datacurve |
| DeepSWE v1.1 | Max | 67.2% | Datacurve |
| DeepSWE v1.1 | Medium | 11.3% | Datacurve |
| DeepSWE v1.1 | XHigh | 56.9% | Datacurve |
| EnterpriseOps-Gym-AA | Max | 40.8% | Artificial Analysis |
| Epoch · Chess Puzzles · v2.0.0 | Low | 17.0% | Epoch AI |
| Epoch · Chess Puzzles · v2.0.0 | Max | 43.0% | Epoch AI |
| Epoch · Chess Puzzles · v2.0.0 | None | 3.0% | Epoch AI |
| Epoch · FrontierMath Tier 4 · v2.0.0 | Max | 61.0% | Epoch AI |
| Epoch · FrontierMath Tiers 1–3 · v2.0.0 | Low | 41.4% | Epoch AI |
| Epoch · FrontierMath Tiers 1–3 · v2.0.0 | Max | 82.1% | Epoch AI |
| Epoch · FrontierMath Tiers 1–3 · v2.0.0 | None | 39.6% | Epoch AI |
| Epoch · GPQA diamond · v1.0.11 | Low | 82.3% | Epoch AI |
| Epoch · GPQA diamond · v1.0.11 | Max | 91.6% | Epoch AI |
| Epoch · GPQA diamond · v1.0.11 | None | 63.6% | Epoch AI |
| Epoch · Mystery Game Puzzles · v1.0.4 · Tool submission | Low | 17.0% | Epoch AI |
| Epoch · Mystery Game Puzzles · v1.0.4 · Tool submission | Max | 21.0% | Epoch AI |
| Epoch · Mystery Game Puzzles · v1.0.4 · Tool submission | Medium | 12.0% | Epoch AI |
| Epoch · Mystery Game Puzzles · v1.0.4 · Tool submission | None | 20.0% | Epoch AI |
| Epoch · OTIS Mock AIME 2024-2025 · v1.1.6 | Low | 66.7% | Epoch AI |
| Epoch · OTIS Mock AIME 2024-2025 · v1.1.6 | Max | 98.3% | Epoch AI |
| Epoch · OTIS Mock AIME 2024-2025 · v1.1.6 | None | 40.0% | Epoch AI |
| Epoch · SimpleQA Verified · v1.0.0 | Max | 41.0% | Epoch AI |
| FrontierCode 1.1 · extended-codex | High | 50.5% | Cognition |
| FrontierCode 1.1 · extended-codex | Low | 29.0% | Cognition |
| FrontierCode 1.1 · extended-codex | Max | 55.1% | Cognition |
| FrontierCode 1.1 · extended-codex | Medium | 41.2% | Cognition |
| FrontierCode 1.1 · extended-codex | XHigh | 54.5% | Cognition |
| FrontierCode 1.1 · main-codex | High | 35.9% | Cognition |
| FrontierCode 1.1 · main-codex | Low | 15.4% | Cognition |
| FrontierCode 1.1 · main-codex | Max | 39.8% | Cognition |
| FrontierCode 1.1 · main-codex | Medium | 25.7% | Cognition |
| FrontierCode 1.1 · main-codex | XHigh | 38.9% | Cognition |
| GDP.pdf: All-pass Rate | High | 22.2% | Artificial Analysis |
| GDP.pdf: All-pass Rate | Low | 14.2% | Artificial Analysis |
| GDP.pdf: All-pass Rate | Max | 24.0% | Artificial Analysis |
| GDP.pdf: All-pass Rate | Medium | 15.2% | Artificial Analysis |
| GDP.pdf: All-pass Rate | Non-reasoning | 6.0% | Artificial Analysis |
| GDP.pdf: All-pass Rate | XHigh | 23.8% | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | High | 1339 | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | Low | 1002 | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | Max | 1470 | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | Medium | 1137 | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | Non-reasoning | 928 | Artificial Analysis |
| GDPval-AA v2.1 Leaderboard | XHigh | 1407 | Artificial Analysis |
| GPQA Diamond | High | 89.2% | Artificial Analysis |
| GPQA Diamond | Low | 83.5% | Artificial Analysis |
| GPQA Diamond | Max | 91.1% | Artificial Analysis |
| GPQA Diamond | Medium | 85.9% | Artificial Analysis |
| GPQA Diamond | Non-reasoning | 64.5% | Artificial Analysis |
| GPQA Diamond | XHigh | 89.5% | Artificial Analysis |
| Humanity's Last Exam | High | 33.4% | Artificial Analysis |
| Humanity's Last Exam | Low | 19.8% | Artificial Analysis |
| Humanity's Last Exam | Max | 39.5% | Artificial Analysis |
| Humanity's Last Exam | Medium | 25.8% | Artificial Analysis |
| Humanity's Last Exam | Non-reasoning | 7.2% | Artificial Analysis |
| Humanity's Last Exam | XHigh | 37.0% | Artificial Analysis |
| Humanity's Sixth Sense | Max | 21.6% | Scale |
| ITBench-AA | Max | 40.3% | Artificial Analysis |
| LiveBench · AMPS Hard · 2026-06-25 | Max | 98.0% | LiveBench |
| LiveBench · code completion · 2026-06-25 | Max | 87.0% | LiveBench |
| LiveBench · code generation · 2026-06-25 | Max | 78.9% | LiveBench |
| LiveBench · connections · 2026-06-25 | Max | 96.5% | LiveBench |
| LiveBench · consecutive events · 2026-06-25 | Max | 86.5% | LiveBench |
| LiveBench · integrals with game · 2026-06-25 | Max | 70.0% | LiveBench |
| LiveBench · javascript · 2026-06-25 | Max | 63.6% | LiveBench |
| LiveBench · logic with navigation · 2026-06-25 | Max | 80.0% | LiveBench |
| LiveBench · math comp · 2026-06-25 | Max | 92.2% | LiveBench |
| LiveBench · olympiad · 2026-06-25 | Max | 88.6% | LiveBench |
| LiveBench · paraphrase · 2026-06-25 | Max | 47.8% | LiveBench |
| LiveBench · plot unscrambling · 2026-06-25 | Max | 51.2% | LiveBench |
| LiveBench · python · 2026-06-25 | Max | 45.0% | LiveBench |
| LiveBench · simplify · 2026-06-25 | Max | 58.7% | LiveBench |
| LiveBench · spatial · 2026-06-25 | Max | 96.0% | LiveBench |
| LiveBench · story generation · 2026-06-25 | Max | 69.1% | LiveBench |
| LiveBench · summarize · 2026-06-25 | Max | 64.9% | LiveBench |
| LiveBench · tablejoin · 2026-06-25 | Max | 47.6% | LiveBench |
| LiveBench · tablereformat · 2026-06-25 | Max | 100.0% | LiveBench |
| LiveBench · theory of mind · 2026-06-25 | Max | 73.1% | LiveBench |
| LiveBench · typescript · 2026-06-25 | Max | 36.7% | LiveBench |
| LiveBench · typos · 2026-06-25 | Max | 70.0% | LiveBench |
| LiveBench · zebra puzzle · 2026-06-25 | Max | 93.5% | LiveBench |
| LMArena Agent Arena | XHigh | -1.2 | LMArena |
| MaxxBench | Max | 46.1% | unLAP |
| MLCR-AA | High | 16.7% | Artificial Analysis |
| MLCR-AA | Low | 15.6% | Artificial Analysis |
| MLCR-AA | Max | 20.0% | Artificial Analysis |
| MLCR-AA | Medium | 13.3% | Artificial Analysis |
| MLCR-AA | Non-reasoning | 8.9% | Artificial Analysis |
| MLCR-AA | XHigh | 15.0% | Artificial Analysis |
| MMMU-Pro | High | 77.6% | Artificial Analysis |
| MMMU-Pro | Low | 74.0% | Artificial Analysis |
| MMMU-Pro | Max | 78.6% | Artificial Analysis |
| MMMU-Pro | Medium | 75.8% | Artificial Analysis |
| MMMU-Pro | Non-reasoning | 60.3% | Artificial Analysis |
| MMMU-Pro | XHigh | 78.6% | Artificial Analysis |
| SciCode | High | 51.6% | Artificial Analysis |
| SciCode | Low | 46.1% | Artificial Analysis |
| SciCode | Max | 53.6% | Artificial Analysis |
| SciCode | Medium | 46.8% | Artificial Analysis |
| SciCode | Non-reasoning | 40.4% | Artificial Analysis |
| SciCode | XHigh | 50.5% | Artificial Analysis |
| SWE-rebench · 2026-06 · tools | Medium | 31.0% | SWE-rebench |
| SWE-sweep | High | 1.4% | SWE-sweep |
| SWE-sweep | Unspecified | 0.5% | SWE-sweep |
| SWE-sweep | XHigh | 2.5% | SWE-sweep |
| Terminal-Bench 2.1 | Max | 75.7% | Terminal-Bench |
| Terminal-Bench 4.0 · codex | Max | 17.3% | Terminal-Bench |
| Terminal-Bench v2.1 | High | 69.7% | Artificial Analysis |
| Terminal-Bench v2.1 | Low | 43.4% | Artificial Analysis |
| Terminal-Bench v2.1 | Max | 80.9% | Artificial Analysis |
| Terminal-Bench v2.1 | Medium | 53.2% | Artificial Analysis |
| Terminal-Bench v2.1 | Non-reasoning | 39.0% | Artificial Analysis |
| Terminal-Bench v2.1 | XHigh | 77.9% | Artificial Analysis |
| Terminal-Bench v4.0 | High | 2.5% | Artificial Analysis |
| Terminal-Bench v4.0 | Low | 0.0% | Artificial Analysis |
| Terminal-Bench v4.0 | Max | 11.6% | Artificial Analysis |
| Terminal-Bench v4.0 | Medium | 0.5% | Artificial Analysis |
| Terminal-Bench v4.0 | Non-reasoning | 1.0% | Artificial Analysis |
| Terminal-Bench v4.0 | XHigh | 3.5% | Artificial Analysis |
| Terminal-Bench-Science 0.1 | Max | 3.3% | Artificial Analysis |
| Vals AI · Code Migration v1 | Max | 44.5% | Vals AI |
| Vals AI · Excel Modeling Benchmark v1 | Max | 67.1% | Vals AI |
| Vals AI · Finance Agent v2 | Max | 55.0% | Vals AI |
| Vals AI · Legal Research Bench v1 | Max | 36.5% | Vals AI |
| Vals AI · MMLU-Pro v1 | Max | 86.0% | Vals AI |
| Vals AI · Tax Agent Bench v1: Weighted Pass | Max | 60.8% | Vals AI |
| Vals AI · Terminal-Bench 2.1 | Max | 79.0% | Vals AI |
| Vals AI · Vibe Code Bench v1.1 | Max | 77.1% | Vals AI |
| 𝜏³-Banking | High | 25.2% | Artificial Analysis |
| 𝜏³-Banking | Low | 12.8% | Artificial Analysis |
| 𝜏³-Banking | Max | 31.1% | Artificial Analysis |
| 𝜏³-Banking | Medium | 17.7% | Artificial Analysis |
| 𝜏³-Banking | Non-reasoning | 9.7% | Artificial Analysis |
| 𝜏³-Banking | XHigh | 28.7% | Artificial Analysis |
Benchmark scores & sources
Original results, evaluation harnesses, and evidence behind this model.
| Benchmark | Bucket | Score | Harness | Evidence | Source-recorded date |
|---|---|---|---|---|---|
| GPQA Diamond · not specified | Hard reasoning | 92.3%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Academic table / GPQA Diamond / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| GraphWalks BFS 1mil f1 · not specified | Long context | 51.2%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified; BFS 1mil f1 Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / GraphWalks BFS 1mil f1 / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| GraphWalks BFS 256k f1 · not specified | Long context | 81.3%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified; BFS 256k f1 Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / GraphWalks BFS 256k f1 / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| HealthBench · not specified | Supporting evidence | 55.8 score (0-100)Reported settings & sourcelength-adjusted; official HealthBench scoring Mean response length 1930 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench / gpt-5.6-luna / length-adjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench · not specified | Supporting evidence | 55.4 score (0-100)Reported settings & sourceunadjusted; official HealthBench scoring Mean response length 1930 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench / gpt-5.6-luna / unadjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench Consensus · not specified | Supporting evidence | 95.1 score (0-100)Reported settings & sourcelength-adjusted; official HealthBench scoring Mean response length 1897 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench Consensus / gpt-5.6-luna / length-adjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench Consensus · not specified | Supporting evidence | 95.1 score (0-100)Reported settings & sourceunadjusted; official HealthBench scoring Mean response length 1897 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench Consensus / gpt-5.6-luna / unadjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench Hard · not specified | Supporting evidence | 32 score (0-100)Reported settings & sourcelength-adjusted; official HealthBench scoring Mean response length 1923 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench Hard / gpt-5.6-luna / length-adjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench Hard · not specified | Supporting evidence | 31.4 score (0-100)Reported settings & sourceunadjusted; official HealthBench scoring Mean response length 1923 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench Hard / gpt-5.6-luna / unadjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench Professional · not specified | Supporting evidence | 55.7%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified; official paper scoring; length-adjusted Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Science And Health table / HealthBench Professional / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench Professional · not specified | Supporting evidence | 55.7 score (0-100)Reported settings & sourcelength-adjusted; official HealthBench scoring Mean response length 3389 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench Professional / gpt-5.6-luna / length-adjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| HealthBench Professional · not specified | Supporting evidence | 59.8 score (0-100)Reported settings & sourceunadjusted; official HealthBench scoring Mean response length 3389 characters. Adjustment centered on2000characters; adjusted and unadjusted values remain distinct configurations. GPT-6 Astra System Card · Table6 / HealthBench Professional / gpt-5.6-luna / unadjusted · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| Humanity's Sixth Sense | Supporting evidence | 21.6% | HSS · maxReported settingsFull 522-task pass@1, task mean of three attempts; every rubric criterion must pass. 95% task-cluster bootstrap CI. Claude-Opus-5 judge. Scale AI + Elorian. Unversioned board; source hashes identify captures. Display only. Excluded from weighted Overall. Source-recorded date is the capture date, not a proven evaluation date. | official board | 2026-10-31 |
| Internal Research Debugging Evaluation · not specified | Supporting evidence | 50.8%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / Internal Research Debugging Evaluation / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| KernelGen 1P · not specified | Supporting evidence | 22.4%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / KernelGen 1P / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| LifeSciBench · not specified | Supporting evidence | 51.2%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Science And Health table / LifeSciBench / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| Management Consulting Tasks (Internal) · not specified | Supporting evidence | 35.4%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Professional table / Management Consulting Tasks (Internal) / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| MedChemBench (Internal) · not specified | Supporting evidence | 30.4%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Science And Health table / MedChemBench (Internal) / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| MMMU Pro (no tools) · not specified | Multimodal | 78.4%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified; no tools Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Multimodal table / MMMU Pro (no tools) / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| MMMU Pro (with tools) · not specified | Multimodal | 79.5%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified; tools enabled Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Multimodal table / MMMU Pro (with tools) / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| NanoGPT · not specified | Supporting evidence | 1.66%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / NanoGPT / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |
| OpenAI MRCR v2 8-needle 256K-512K · v2 | Long context | 41.3%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified; 8-needle 256K-512K Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / OpenAI MRCR v2 8-needle 256K-512K / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| OpenAI MRCR v2 8-needle 512K-1M · v2 | Long context | 41.3%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified; 8-needle 512K-1M Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Long Context table / OpenAI MRCR v2 8-needle 512K-1M / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| OSWorld 2.0 · 2.0 | Agentic | 45.6%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Computer Use table / OSWorld 2.0 / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configurationcontributes to capability | lab self-report | Reviewed 2026-09-06 |
| PostTrainBench Lite · not specified | Supporting evidence | 29.6%Reported settings & sourceLaunch table reported configuration; per-cell reasoning effort unspecified Provider-published result; comparator measurements are not automatically independently reproduced. GPT-5.6: Frontier intelligence that scales with your ambition · Self-Improvement table / PostTrainBench Lite / GPT‑5.6 Luna · reviewed 2026-09-06 | Published configuration | lab self-report | Reviewed 2026-09-06 |