# GPT-5.6 LunaMedium

> Source: Benchize (https://benchize.com) — https://benchize.com/models/gpt-5-6-luna-medium. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/models/gpt-5-6-luna-medium.

OpenAI

Canonical URL: https://benchize.com/models/gpt-5-6-luna-medium
Last updated: Thu, 22 Oct 2026 09:10:07 GMT

Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations.

OpenAI

## Benchmarks

| Benchmark | Score | ± | Rank | Cost | Latency | Source |  |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [ARC-AGI-1](https://benchize.com/benchmarks/arc-prize%3Av1_Semi_Private.md) | 56.5% | — | 158 / 245 | 0.06usd per task | — | [arc prize · v1](https://arcprize.org/media/data/leaderboard/v1.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.565 \| score_fraction \| v1_Semi_Private \| Higher is better \| \| `cost_per_task` \| 0.06 \| usd_per_task \| v1_Semi_Private \| Lower is better \| ### Configuration evaluationSplit v1_Semi_Private harness Not reported modelGroup openai-gpt-5-6-luna modelType CoT Reasoning effort medium sourceModelVersion openai-gpt-5-6-luna-medium thinkingEnabled true ### Source [Original source](https://arcprize.org/media/data/leaderboard/v1.json) Source arc prize Original model label GPT-5.6 Luna (Medium) Source version v1 Observed 2026-10-22T05:20:04.066Z Evidence SHA-256 `2bf68d22dafba37047d07ad833ebc0e8199c9aa2f9eeaed50c75c1b704462836` Source locator `/evaluations/165` |  |  |  |  |  |  |  |
| [ToneBench Scriptwriting](https://benchize.com/benchmarks/tonebench%3Ascripts.md) | 1,592.4 | Range 1,550.3–1,632.7 | 71 / 126 | — | — | [tonebench](https://benchmark.towardsai.com/data/leaderboard.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `elo` \| 1592.4 \| rating \| reported \| Higher is better \| \| `overall` \| 76.66 \| percent \| reported \| Higher is better \| \| `anti_slop_numbers` \| 71.18 \| source_score \| rubric \| Higher is better \| \| `continuity_emotion` \| 77.06 \| source_score \| rubric \| Higher is better \| \| `cue_quality` \| 76.1 \| source_score \| rubric \| Higher is better \| \| `hook_strength` \| 82.25 \| source_score \| rubric \| Higher is better \| \| `length_adherence` \| 70.86 \| source_score \| rubric \| Higher is better \| \| `slop` \| 82.95 \| source_score \| rubric \| Higher is better \| \| `slop_score` \| 94.72 \| source_score \| rubric \| Higher is better \| \| `substance_accuracy` \| 83.61 \| source_score \| rubric \| Higher is better \| \| `tone_voice` \| 73.24 \| source_score \| rubric \| Higher is better \| \| `writing_quality` \| 75.55 \| source_score \| rubric \| Higher is better \| \| `youtube_best_practices` \| 70.92 \| source_score \| rubric \| Higher is better \| ### Configuration provider codex Reasoning effort unknown scoreScope 10_task_current sourceConfiguration gpt-5-6-luna thinkingEnabled true ### Reported statistics Uncertainty Reported interval: 1,550.3–1,632.7 interval [1550.3,1632.7] rank 72 runCount 5 ### Source [Original source](https://benchmark.towardsai.com/data/leaderboard.json) Source tonebench Original model label GPT-5.6 Luna (medium) Evaluated 2026-10-11 Observed 2026-10-22T04:40:14.102Z Evidence SHA-256 `4264338337f418d3ded7ace250cad64fe12cacbd4b79a312bfaf0ae9268a01c3` Source locator `/rows/71` |  |  |  |  |  |  |  |
| [FrontierCode 1.1 main](https://benchize.com/benchmarks/frontiercode%3Av1_1%3Amain.md) | 25.74% | — | 107 / 125 | $0.13 | — | [frontiercode](https://cognition.com/data/frontiercode-leaderboard/data.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.2574 \| score_fraction \| blocking_criteria_adjusted \| Higher is better \| \| `correct` \| 0.2909 \| score_fraction \| raw_rubric \| Higher is better \| \| `cost` \| 0.1275 \| USD \| per_rollout \| Lower is better \| ### Configuration harness codex Reasoning effort medium sourceVersion v1_1 subset main ### Reported statistics flaggedRate 0 sampleSize 100 ### Source [Original source](https://cognition.com/data/frontiercode-leaderboard/data.json) Source frontiercode Original model label GPT-5.6 Luna Observed 2026-10-22T05:40:03.220Z Evidence SHA-256 `ff0040778b7bdf69c705ad7474f771fb5e53cfaf5cf8573b3a4614153543d75f` Source locator `/v1_1/data/GPT-5.6%20Luna/medium/main` |  |  |  |  |  |  |  |
| [FrontierCode 1.1 extended](https://benchize.com/benchmarks/frontiercode%3Av1_1%3Aextended.md) | 41.17% | — | 107 / 125 | $0.12 | — | [frontiercode](https://cognition.com/data/frontiercode-leaderboard/data.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.4117 \| score_fraction \| blocking_criteria_adjusted \| Higher is better \| \| `correct` \| 0.4566 \| score_fraction \| raw_rubric \| Higher is better \| \| `cost` \| 0.115 \| USD \| per_rollout \| Lower is better \| ### Configuration harness codex Reasoning effort medium sourceVersion v1_1 subset extended ### Reported statistics flaggedRate 0 sampleSize 150 ### Source [Original source](https://cognition.com/data/frontiercode-leaderboard/data.json) Source frontiercode Original model label GPT-5.6 Luna Observed 2026-10-22T05:40:03.220Z Evidence SHA-256 `ff0040778b7bdf69c705ad7474f771fb5e53cfaf5cf8573b3a4614153543d75f` Source locator `/v1_1/data/GPT-5.6%20Luna/medium/extended` |  |  |  |  |  |  |  |
| [ARC-AGI-2](https://benchize.com/benchmarks/arc-prize%3Av2_Semi_Private.md) | 7.36% | — | 153 / 243 | 0.11usd per task | — | [arc prize · v2](https://arcprize.org/media/data/leaderboard/v2.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.0736 \| score_fraction \| v2_Semi_Private \| Higher is better \| \| `cost_per_task` \| 0.11 \| usd_per_task \| v2_Semi_Private \| Lower is better \| ### Configuration evaluationSplit v2_Semi_Private harness Not reported modelGroup openai-gpt-5-6-luna modelType CoT Reasoning effort medium sourceModelVersion openai-gpt-5-6-luna-medium thinkingEnabled true ### Source [Original source](https://arcprize.org/media/data/leaderboard/v2.json) Source arc prize Original model label GPT-5.6 Luna (Medium) Source version v2 Observed 2026-10-22T05:20:04.306Z Evidence SHA-256 `4351eb696e78113b0459e9bb5f0258318b905efef34f473dcea51ce641b0d94d` Source locator `/evaluations/166` |  |  |  |  |  |  |  |
| [Mystery Game Puzzles](https://benchize.com/benchmarks/epoch-ai%3Amystery_game_puzzles.md) | 12% | SE ±3.27 pp | 59 / 72 | — | — | [epoch ai · Mystery Game Puzzles](https://epoch.ai/data/benchmark_data.zip#file=mystery_game_puzzles.csv) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.12 \| score_fraction \| Best score (across scorers) \| Higher is better \| \| `mean_score` \| 0.12 \| score_fraction \| mean_score \| Higher is better \| ### Configuration benchmarkVersion Mystery Game Puzzles reasoningBudget Not reported Reasoning effort medium sourceModelVersion gpt-5.6-luna_medium ### Reported statistics Uncertainty Standard error: ±3.27 pp standardError 0.032659863237109045 ### Source [Original source](https://epoch.ai/data/benchmark_data.zip#file=mystery_game_puzzles.csv) Source epoch ai Original model label gpt-5.6-luna Source version Mystery Game Puzzles Evaluated 2026-08-27T18:48:42.000Z Observed 2026-10-22T03:00:05.281Z Evidence SHA-256 `f66941ed5e9520190b66af57d2b6fd62a9582077dd25ab8764ce651bc0eca00a` Source locator `/rows/46` |  |  |  |  |  |  |  |
| [NC Bench](https://benchize.com/benchmarks/nc-bench%3Aoverall.md) | 90.61% | — | 54 / 144 | — | — | [nc bench](https://www.nc-bench.com/models) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 90.61 \| percent \| reported \| Higher is better \| \| `tooling` \| 99.97 \| percent \| category \| Higher is better \| \| `creative_writing` \| 90.1 \| percent \| category \| Higher is better \| \| `language` \| 96.67 \| percent \| category \| Higher is better \| \| `utility` \| 92.18 \| percent \| category \| Higher is better \| \| `reasoning` \| 84.74 \| percent \| category \| Higher is better \| \| `text_editing` \| 97.77 \| percent \| category \| Higher is better \| \| `rule_following` \| 64.83 \| percent \| category \| Higher is better \| \| `hallucination` \| 98.65 \| percent \| category \| Higher is better \| ### Configuration Reasoning effort medium Model configuration GPT-5.6 Luna (Reasoning, Medium) thinkingEnabled true ### Source [Original source](https://www.nc-bench.com/models) Source nc bench Original model label GPT-5.6 Luna (Reasoning, Medium) Observed 2026-10-22T06:10:03.426Z Evidence SHA-256 `ec261022458543624b08a136d01f5a2f0666cd52ffb49c64d6ca983e8543d1b5` Source locator `/rows/66` |  |  |  |  |  |  |  |
| [DeepSWE](https://benchize.com/benchmarks/deepswe%3Av1.1.md) | 11.28% | Range 10.45–12.11% | 69 / 70 | $0.22 | 181.53s | [deepswe](https://deepswe.datacurve.ai/artifacts/v1.1/leaderboard-live.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `pass_at_1` \| 0.11283185840707964 \| score_fraction \| scored_rollout_attempts \| Higher is better \| \| `pass_at_4` \| 0.2743362831858407 \| score_fraction \| tasks_with_one_passing_rollout \| Higher is better \| \| `cost_per_attempt` \| 0.21630978982300886 \| USD \| scored_attempt \| Lower is better \| \| `duration` \| 181.5252541438053 \| seconds \| scored_attempt \| Lower is better \| ### Configuration harness mini-swe-agent Reasoning effort medium sourceConfiguration mini_swe_agent_gpt_5_6_luna_medium ### Reported statistics Uncertainty Reported interval: 10.45–12.11% interval [0.10452866084502313,0.12113505596913615] intervalMethod 95% run-to-run: SE across repeated whole-benchmark passes (1.96 * std(runs)/sqrt(R)) runCount 4 sampleSize 452 taskCount 113 ### Source [Original source](https://deepswe.datacurve.ai/artifacts/v1.1/leaderboard-live.json) Source deepswe Original model label gpt-5-6-luna Observed 2026-10-22T03:50:05.889Z Evidence SHA-256 `a7c15d66288fd249c020b9931c017b92d1a3b90e480b3ff34974b752bd030019` Source locator `/rows/68` |  |  |  |  |  |  |  |
| [ARC-AGI-3](https://benchize.com/benchmarks/arc-prize%3Av3_Semi_Private.md) | 0.17% | — | 55 / 65 | 2,379.13usd | — | [arc prize · v3](https://arcprize.org/media/data/leaderboard/v3.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.0017 \| score_fraction \| v3_Semi_Private \| Higher is better \| \| `total_cost` \| 2379.13 \| usd \| v3_Semi_Private \| Lower is better \| ### Configuration evaluationSplit v3_Semi_Private harness standard modelGroup openai-gpt-5-6-luna modelType CoT Reasoning effort medium sourceModelVersion openai-gpt-5-6-luna-medium thinkingEnabled true ### Source [Original source](https://arcprize.org/media/data/leaderboard/v3.json) Source arc prize Original model label GPT-5.6 Luna (Medium) Source version v3 Observed 2026-10-22T05:20:04.578Z Evidence SHA-256 `3ce335487296bb2dcf4ba9546c5a2a02b1907db5b08887cd9ff4429fed824916` Source locator `/evaluations/18` |  |  |  |  |  |  |  |
| [AutomationBench](https://benchize.com/benchmarks/zapier%3Aautomationbench.md) | 4.26% | — | 90 / 113 | $0.02 | — | [zapier](https://framerusercontent.com/sites/4WTSl4BNjd1q9QFEFibC6h/MBpMAlj5k_Qpj1iTRcXO7kAjUP_VB-QuD3gMvxxThCQ.C6QAp2jN.mjs) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 4.26 \| percent \| reported \| Higher is better \| \| `cost` \| 0.02 \| USD \| per_task \| Lower is better \| ### Configuration benchmarkVersion 1.0.6 costFootnote false Fallback policy Not reported Reasoning effort medium Model configuration GPT-5.6 Luna (Medium) thinkingEnabled true ### Reported statistics rank 102 ### Source [Original source](https://framerusercontent.com/sites/4WTSl4BNjd1q9QFEFibC6h/MBpMAlj5k_Qpj1iTRcXO7kAjUP_VB-QuD3gMvxxThCQ.C6QAp2jN.mjs) Source zapier Original model label GPT-5.6 Luna (Medium) Observed 2026-10-22T08:20:03.209Z Evidence SHA-256 `6f5ef9e0d3799e2bac5ae924407ac89f608f18c17f3ef238887356808b094579` Source locator `/rows/101` |  |  |  |  |  |  |  |
| [SWE-rebench](https://benchize.com/benchmarks/swe-rebench%3Aresolved.md) | 43.6% | — | 4 / 8 | $0.11 | — | [swe rebench](https://swe-rebench.com/) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `resolved` \| 43.6 \| percent \| reported \| Higher is better \| \| `pass_at_5` \| 59.5 \| percent \| reported \| Higher is better \| \| `cost_per_problem` \| 0.11 \| USD \| per_problem \| Lower is better \| ### Configuration Reasoning effort medium Model configuration GPT-5.6 Luna [medium] submissionType model thinkingEnabled true ### Source [Original source](https://swe-rebench.com/) Source swe rebench Original model label GPT-5.6 Luna [medium] Observed 2026-10-22T06:00:03.589Z Evidence SHA-256 `fbb1b659562843da266aa20ec2295253d727ccd9b9469c78ab32df6fad1fffed` Source locator `/rows/12` |  |  |  |  |  |  |  |

---
Cite as: GPT-5.6 LunaMedium — Benchize, https://benchize.com/models/gpt-5-6-luna-medium (retrieved 2026-10-22)