Model comparison
GPT-5.5 vs GPT-5.6 Luna
Head-to-head evidence from 36 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: GPT-5.5 #9 (Estimated); GPT-5.6 Luna #22 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.5 and GPT-5.6 Luna share 36 comparable benchmark results. 5 of 8 categories are comparable. 21 results are unique to GPT-5.5; 5 to GPT-5.6 Luna.
Updated July 23, 2026- Shared results
- 36
- GPT-5.5 only
- 21
- GPT-5.6 Luna only
- 5
- Comparable categories
- 5 / 8
Pick GPT-5.5 if you want the stronger benchmark profile. GPT-5.6 Luna only becomes the better choice if knowledge is the priority or you want the cheaper token bill.
Confidence note. This is a partial-evidence comparison with 36 shared benchmark results across 6 evidence categories; 5 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
GPT-5.5 is clearly ahead on the BenchAlign aggregate, 73.51 to 67.17. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.5 is also the more expensive model on tokens at $5.00 input / $30.00 output per 1M tokens, versus $1.00 input / $6.00 output per 1M tokens for GPT-5.6 Luna. That is roughly 5.0x on output cost alone.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | GPT-5.5 | Δ | GPT-5.6 Luna |
|---|---|---|---|
| Knowledge | GPT-5.557.8 | Margin→ 34.5 | GPT-5.6 Luna92.3 |
| Math | GPT-5.547.6 | Margin→ 26.0 | GPT-5.6 Luna73.6 |
| Multimodal | GPT-5.570.4 | Margin→ 8.0 | GPT-5.6 Luna78.4 |
| Coding | GPT-5.558.6 | Margin→ 4.1 | GPT-5.6 Luna62.7 |
| Agentic | GPT-5.581.6 | Margin→ 2.5 | GPT-5.6 Luna84.1 |
| Reasoning | GPT-5.585.0 | MarginNo overlap | GPT-5.6 LunaNot measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
FrontierMath v2 (Tiers 1-3)
MathA 51.700%B 78.600%Winner: GPT-5.6 LunaΔ 26.9FrontierMath v2 (Tiers 1-3): GPT-5.5 scored 51.700%; GPT-5.6 Luna scored 78.600%. GPT-5.6 Luna wins this benchmark. - Source ↗
FrontierMath v2 (Tier 4)
MathA 35.400%B 58.500%Winner: GPT-5.6 LunaΔ 23.1FrontierMath v2 (Tier 4): GPT-5.5 scored 35.400%; GPT-5.6 Luna scored 58.500%. GPT-5.6 Luna wins this benchmark. - Source ↗
SWE-bench Pro
CodingA 58.6%B 62.7%Winner: GPT-5.6 LunaΔ 4.1SWE-bench Pro: GPT-5.5 scored 58.6%; GPT-5.6 Luna scored 62.7%. GPT-5.6 Luna wins this benchmark. - Source ↗
MMMU-Pro
MultimodalA 81.2%B 78.4%Winner: GPT-5.5Δ 2.8MMMU-Pro: GPT-5.5 scored 81.2%; GPT-5.6 Luna scored 78.4%. GPT-5.5 wins this benchmark. - Source ↗
Terminal-Bench 2.0
AgenticA 82%B 84.7%Winner: GPT-5.6 LunaΔ 2.7Terminal-Bench 2.0: GPT-5.5 scored 82%; GPT-5.6 Luna scored 84.7%. GPT-5.6 Luna wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.5 | GPT-5.6 Luna | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.5$5 input / $30 output | GPT-5.6 Luna$1 input / $6 output | GPT-5.6 Luna has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.5Not available | GPT-5.6 LunaNot available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | GPT-5.5Not available | GPT-5.6 LunaNot available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | GPT-5.51M | GPT-5.6 Luna1M | Listed context windows are equal. |
Benchmark Deep Dive
AgenticGPT-5.6 Luna wins24 benchmarks
| Benchmark | GPT-5.5 | GPT-5.6 Luna | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 82% | 84.7% | GPT-5.6 Luna leads |
| CyberGymSource | 81.8% | 77.9% | GPT-5.5 leads |
| BrowseCompSource | 84.4% | 83.3% | GPT-5.5 leads |
| OSWorld-VerifiedSource | 78.7% | — | Not comparable |
| MCP AtlasSource | 75.3% | — | Not comparable |
| ToolathlonSource | 55.6% | 53.4% | GPT-5.5 leads |
| τ²-bench resultsSource | 93.9% | — | Not comparable |
| AA Agentic IndexSource | 44.9% | 45.6% | GPT-5.6 Luna leads |
| APEX-Agents-AASource | 37.7% | 35.8% | GPT-5.5 leads |
| GDPval-AASource | 49.5% | 54.2% | GPT-5.6 Luna leads |
| GDPval-AASource | 1490 | 1584 | GPT-5.6 Luna leads |
| Gert LabsSource | 72.93% | — | Not comparable |
| ResearchClawBenchSource | 17.0% | — | Not comparable |
| OSWorld 2.0Source | 13.0% | 45.6% | GPT-5.6 Luna leads |
| JobBenchSource | 42.7% | — | Not comparable |
| ExploitGymSource | 13.4% | 12.4% | GPT-5.5 leads |
| AA BriefcaseSource | 1154 | — | Not comparable |
| AA AutomationBenchSource | 42.1% | 42.2% | GPT-5.6 Luna leads |
| AA EnterpriseOps-GymSource | 46.6% | — | Not comparable |
| AA Harvey LABSource | 86.3% | 87.9% | GPT-5.6 Luna leads |
| AA ITBenchSource | 45.8% | 40.3% | GPT-5.5 leads |
| AA Tau3 BankingSource | 31.3% | 27.2% | GPT-5.5 leads |
| terminalBenchHardSource | 60.6% | — | Not comparable |
| aaTerminalBench21Source | 84.3% | 80.9% | GPT-5.5 leads |
CodingGPT-5.6 Luna wins11 benchmarks
| Benchmark | GPT-5.5 | GPT-5.6 Luna | Result |
|---|---|---|---|
| SWE-bench ProSource | 58.6% | 62.7% | GPT-5.6 Luna leads |
| Terminal-Bench 2.0Source | 82.0% | 84.7% | GPT-5.6 Luna leads |
| Vibe Code BenchSource | 69.85% | — | Not comparable |
| React Native EvalsSource | 84.7% | — | Not comparable |
| cursorBench31Source | 59.2% | — | Not comparable |
| cursorBench32Source | 58.4% | 61.1% | GPT-5.6 Luna leads |
| AA Coding IndexSource | 74.9% | 71.5% | GPT-5.5 leads |
| AA-SciCodeSource | 56.1% | 52.5% | GPT-5.5 leads |
| FrontierCode 1.1 MainSource | 43.0% | — | Not comparable |
| deepSweSource | — | 67.2% | Not comparable |
| FrontierCode 1.1 ExtendedSource | — | 55.1% | Not comparable |
Reasoning6 benchmarks
KnowledgeGPT-5.6 Luna wins12 benchmarks
| Benchmark | GPT-5.5 | GPT-5.6 Luna | Result |
|---|---|---|---|
| GPQASource | 93.6% | 92.3% | GPT-5.5 leads |
| GPQA-DSource | 93.6% | 92.3% | GPT-5.5 leads |
| HLESource | 52.2% | — | Not comparable |
| HLE w/o toolsSource | 41.4% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 54.8% | 51.2% | GPT-5.5 leads |
| AA-GPQA DiamondSource | 93.5% | 91.1% | GPT-5.5 leads |
| AA-HLESource | 44.3% | 37.2% | GPT-5.5 leads |
| AA-Omniscience IndexSource | 20.1% | -11.2% | GPT-5.5 leads |
| AA-Omniscience AccuracySource | 56.9% | 41.5% | GPT-5.5 leads |
| AA-Omniscience Hallucination RateSource | 85.5% | 90.1% | GPT-5.5 leads |
| HealthBench ProfessionalSource | — | 55.7% | Not comparable |
| HealthBench HardSource | — | 32.0% | Not comparable |
MathGPT-5.6 Luna wins3 benchmarks
MultimodalGPT-5.6 Luna wins5 benchmarks
Inst. Following1 benchmarks
| Benchmark | GPT-5.5 | GPT-5.6 Luna | Result |
|---|---|---|---|
| AA-IFBenchSource | 75.9% | — | Not comparable |
Frequently Asked Questions (6)
Which is better, GPT-5.5 or GPT-5.6 Luna?
GPT-5.5 is ahead on BenchLM's BenchAlign leaderboard, 73.51 to 67.17. The biggest single separator in this matchup is FrontierMath v2 (Tiers 1-3), where the scores are 51.700% and 78.600%.
Which is better for knowledge tasks, GPT-5.5 or GPT-5.6 Luna?
GPT-5.6 Luna has the edge for knowledge tasks in this comparison, averaging 92.3 versus 57.8. Inside this category, AA-Omniscience Index is the benchmark that creates the most daylight between them.
Which is better for coding, GPT-5.5 or GPT-5.6 Luna?
GPT-5.6 Luna has the edge for coding in this comparison, averaging 62.7 versus 58.6. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.
Which is better for math, GPT-5.5 or GPT-5.6 Luna?
GPT-5.6 Luna has the edge for math in this comparison, averaging 73.6 versus 47.6. Inside this category, FrontierMath (legacy) is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, GPT-5.5 or GPT-5.6 Luna?
GPT-5.6 Luna has the edge for agentic tasks in this comparison, averaging 84.1 versus 81.6. Inside this category, GDPval-AA is the benchmark that creates the most daylight between them.
Which is better for multimodal and grounded tasks, GPT-5.5 or GPT-5.6 Luna?
GPT-5.6 Luna has the edge for multimodal and grounded tasks in this comparison, averaging 78.4 versus 70.4. Inside this category, MMMU-Pro w/ Python is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.