# GPT-5.5 vs GPT-5.6 Luna > Side-by-side benchmark comparison for GPT-5.5 and GPT-5.6 Luna across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks. - Canonical page: https://benchlm.ai/compare/gpt-5-5-vs-gpt-5-6-luna - Last updated: September 24, 2026 - Shared sourced benchmarks: 21 - HTML indexing: indexable - Ranking lane: BenchAlign v5.7 ## Quick Verdict Pick GPT-5.5 if you want the stronger benchmark profile. GPT-5.6 Luna only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score. ## Summary - GPT-5.5 leads overall 69.13 to 65.6. - The clearest category separation is in mathematics, where the averages are 69.4 for GPT-5.5 and 94.2 for GPT-5.6 Luna. - The biggest single benchmark swing is OSWorld 2.0 in Agentic, with scores of 13.0% and 45.6%. - GPT-5.6 Luna is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use. - GPT-5.6 Luna also has the larger context window at 1.05M. ## Model Snapshot | Property | GPT-5.5 | GPT-5.6 Luna | |----------|----------|----------| | Creator | OpenAI | OpenAI | | Type | Proprietary | Proprietary | | Reasoning | Reasoning | Reasoning | | Context | 1M | 1.05M | | Overall Score | 69.13 | 65.6 | | Benchmarks Covered | 39 | 29 | | Pricing (input/output) | $5.00 / $30.00 | $0.20 / $1.20 | ## Category Breakdown ### Agentic - Winner: GPT-5.5 - GPT-5.5 public-lane score: 60.1 (Supported · #15/105) - GPT-5.6 Luna public-lane score: 55.3 (Supported · #28/105) | Benchmark | GPT-5.5 | GPT-5.6 Luna | Winner | |-----------|-----------|-----------|--------| | Terminal-Bench 2.0 | 82% | Coming soon | Coming soon | | CyberGym | 81.8% | 77.9% | GPT-5.5 | | BrowseComp | 84.4% | 83.3% | GPT-5.5 | | OSWorld-Verified | 78.7% | Coming soon | Coming soon | | MCP Atlas | 75.3% | Coming soon | Coming soon | | Toolathlon | 55.6% | 53.4% | GPT-5.5 | | τ²-bench results | 98% | Coming soon | Coming soon | | Gert Labs | 72.93% | Coming soon | Coming soon | | ResearchClawBench | 17.0% | Coming soon | Coming soon | | OSWorld 2.0 | 13.0% | 45.6% | GPT-5.6 Luna | | JobBench | 42.7% | Coming soon | Coming soon | | ExploitGym | 13.4% | 12.4% | GPT-5.5 | | Terminal-Bench 2.1 (Vals) | 76.4% | 79.0% | GPT-5.6 Luna | | ApprenticeBench | 20% | 7% | GPT-5.5 | | Terminal-Bench 3.0 | Coming soon | 14.3% | Coming soon | | Terminal-Bench 2.1 | Coming soon | 84.7% | Coming soon | ### Coding - Winner: Tie - GPT-5.5 public-lane score: 64.1 (Supported · #11/135) - GPT-5.6 Luna public-lane score: 64.5 (Supported · #9/135) | Benchmark | GPT-5.5 | GPT-5.6 Luna | Winner | |-----------|-----------|-----------|--------| | SWE-bench Pro | 58.6% | 62.7% | GPT-5.6 Luna | | Terminal-Bench 2.0 | 82.0% | Coming soon | Coming soon | | Vibe Code Bench | 69.85% | Coming soon | Coming soon | | React Native Evals | 84.7% | Coming soon | Coming soon | | cursorBench31 | 59.2% | Coming soon | Coming soon | | cursorBench32 | 58.4% | 61.1% | GPT-5.6 Luna | | FrontierCode 1.1 Main | 43.0% | Coming soon | Coming soon | | LiveCodeBench (Vals) | 85.3% | Coming soon | Coming soon | | SWE-bench (Vals) | 82.6% | 93.0% | GPT-5.6 Luna | | Terminal-Bench 2.1 | Coming soon | 84.7% | Coming soon | | DeepSWE | Coming soon | 67.2% | Coming soon | | FrontierCode 1.1 Extended | Coming soon | 55.1% | Coming soon | | VulcanBench v3 | Coming soon | 85.5% | Coming soon | ### Reasoning - Winner: GPT-5.5 - GPT-5.5 public-lane score: 65.9 (#10/19) - GPT-5.6 Luna public-lane score: 54.7 (#18/19) | Benchmark | GPT-5.5 | GPT-5.6 Luna | Winner | |-----------|-----------|-----------|--------| | MRCR v2 64K-128K | 83.1% | Coming soon | Coming soon | | MRCR v2 128K-256K | 87.5% | Coming soon | Coming soon | | ARC-AGI-2 | 85% | 59.5% | GPT-5.5 | | ARC-AGI-3 | 0.4% | 0.2% | GPT-5.5 | ### Multimodal & Grounded - Winner: Directional only - GPT-5.5 public-lane score: 71.4 (#20/50) - GPT-5.6 Luna public-lane score: 67.1 (#22/50) | Benchmark | GPT-5.5 | GPT-5.6 Luna | Winner | |-----------|-----------|-----------|--------| | MMMU-Pro | 81.2% | 78.4% | GPT-5.5 | | MMMU-Pro w/ Python | 83.2% | 79.5% | GPT-5.5 | | OfficeQA Pro | 54.1% | Coming soon | Coming soon | ### Knowledge - Winner: GPT-5.5 - GPT-5.5 public-lane score: 70 (Supported · #12/158) - GPT-5.6 Luna public-lane score: 64.6 (Supported · #22/158) | Benchmark | GPT-5.5 | GPT-5.6 Luna | Winner | |-----------|-----------|-----------|--------| | GPQA | 93.6% | 92.3% | GPT-5.5 | | GPQA-D | 93.6% | 92.3% | GPT-5.5 | | HLE | 52.2% | Coming soon | Coming soon | | HLE w/o tools | 41.4% | Coming soon | Coming soon | | GPQA Diamond (Vals) | 93.2% | 91.7% | GPT-5.5 | | MMLU-Pro (Vals) | 88.1% | 86.0% | GPT-5.5 | | HealthBench Professional | Coming soon | 55.7% | Coming soon | | HealthBench Hard | Coming soon | 32.0% | Coming soon | ### Instruction Following - Winner: Not comparable - GPT-5.5 public-lane score: 91.9 (#7/124) - GPT-5.6 Luna public-lane score: Coming soon ### Mathematics - Winner: Not comparable - GPT-5.5 public-lane score: 69.4 (Unranked · 3 rankable rows) - GPT-5.6 Luna public-lane score: 94.2 (Unranked · 3 rankable rows) | Benchmark | GPT-5.5 | GPT-5.6 Luna | Winner | |-----------|-----------|-----------|--------| | FrontierMath (legacy) | 51.7% | 78.6% | GPT-5.6 Luna | | FrontierMath v2 (Tiers 1-3) | 51.700% | 78.600% | GPT-5.6 Luna | | FrontierMath v2 (Tier 4) | 35.400% | 58.500% | GPT-5.6 Luna | ## FAQ ### Which is better overall, GPT-5.5 or GPT-5.6 Luna? GPT-5.5 is ahead overall on BenchLM right now. ### Where is the biggest gap between GPT-5.5 and GPT-5.6 Luna? The widest category gap is in mathematics, where the averages are 69.4 for GPT-5.5 and 94.2 for GPT-5.6 Luna. ### How many benchmarks does GPT-5.5 cover on BenchLM? GPT-5.5 currently has 39 sourced benchmark scores on BenchLM. ### How many benchmarks does GPT-5.6 Luna cover on BenchLM? GPT-5.6 Luna currently has 29 sourced benchmark scores on BenchLM. ## Related Comparisons - [GPT-5.5 vs GPT-5.5 Pro](/compare/gpt-5-5-vs-gpt-5-5-pro) - [GPT-5.6 Luna vs GPT-5.5 Pro](/compare/gpt-5-5-pro-vs-gpt-5-6-luna) - [GPT-5.5 vs GPT-5.6 Sol](/compare/gpt-5-5-vs-gpt-5-6-sol) - [GPT-5.6 Luna vs GPT-5.6 Sol](/compare/gpt-5-6-luna-vs-gpt-5-6-sol) ## Explore More - [GPT-5.5 profile](/models/gpt-5-5) - [GPT-5.6 Luna profile](/models/gpt-5-6-luna) - [Compare Pricing](/llm-pricing) - [Alternative Finder](/tools/alternative-finder) - [LLM Selector](/tools/llm-selector) - [Overall Rankings](/best/overall)