Model comparison
GPT-5.6 Luna vs GPT-5.6 Sol
Head-to-head evidence from 40 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Sibling matchup inside the GPT-5.6 family.
Public leaderboard positions: GPT-5.6 Luna #22 (Estimated); GPT-5.6 Sol #3 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.6 Luna and GPT-5.6 Sol share 40 comparable benchmark results. 5 of 8 categories are comparable. 1 result is unique to GPT-5.6 Luna; 6 to GPT-5.6 Sol.
Updated July 23, 2026- Shared results
- 40
- GPT-5.6 Luna only
- 1
- GPT-5.6 Sol only
- 6
- Comparable categories
- 5 / 8
GPT-5.6 Luna makes more sense if you want the cheaper token bill, while GPT-5.6 Sol is the cleaner fit if mathematics is the priority.
Confidence note. This is a partial-evidence comparison with 40 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-5.6 Luna and GPT-5.6 Sol sit in the same GPT-5.6 family. This page is less about two unrelated model lineages and more about how the siblings trade off on benchmark shape, token costs, and practical limits like context window.
GPT-5.6 Sol is clearly ahead on the BenchAlign aggregate, 81.96 to 67.17. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.6 Sol's sharpest advantage is in mathematics, where it averages 87.5 against 73.6. The single biggest benchmark swing on the page is FrontierMath v2 (Tier 4), 58.500% to 83.000%.
GPT-5.6 Sol 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-5.6 Luna | Δ | GPT-5.6 Sol |
|---|---|---|---|
| Math | GPT-5.6 Luna73.6 | Margin→ 13.9 | GPT-5.6 Sol87.5 |
| Agentic | GPT-5.6 Luna84.1 | Margin→ 7.9 | GPT-5.6 Sol92.0 |
| Multimodal | GPT-5.6 Luna78.4 | Margin→ 4.6 | GPT-5.6 Sol83.0 |
| Knowledge | GPT-5.6 Luna92.3 | Margin→ 2.3 | GPT-5.6 Sol94.6 |
| Coding | GPT-5.6 Luna62.7 | Margin→ 1.9 | GPT-5.6 Sol64.6 |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
FrontierMath v2 (Tier 4)
MathA 58.500%B 83.000%Winner: GPT-5.6 SolΔ 24.5FrontierMath v2 (Tier 4): GPT-5.6 Luna scored 58.500%; GPT-5.6 Sol scored 83.000%. GPT-5.6 Sol wins this benchmark. - Source ↗
FrontierMath v2 (Tiers 1-3)
MathA 78.600%B 89.000%Winner: GPT-5.6 SolΔ 10.4FrontierMath v2 (Tiers 1-3): GPT-5.6 Luna scored 78.600%; GPT-5.6 Sol scored 89.000%. GPT-5.6 Sol wins this benchmark. - Source ↗
BrowseComp
AgenticA 83.3%B 92.2%Winner: GPT-5.6 SolΔ 8.9BrowseComp: GPT-5.6 Luna scored 83.3%; GPT-5.6 Sol scored 92.2%. GPT-5.6 Sol wins this benchmark. - Source ↗
Terminal-Bench 2.0
AgenticA 84.7%B 91.9%Winner: GPT-5.6 SolΔ 7.2Terminal-Bench 2.0: GPT-5.6 Luna scored 84.7%; GPT-5.6 Sol scored 91.9%. GPT-5.6 Sol wins this benchmark. - Source ↗
MMMU-Pro
MultimodalA 78.4%B 83%Winner: GPT-5.6 SolΔ 4.6MMMU-Pro: GPT-5.6 Luna scored 78.4%; GPT-5.6 Sol scored 83%. GPT-5.6 Sol wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.6 Luna | GPT-5.6 Sol | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.6 Luna$1 input / $6 output | GPT-5.6 Sol$5 input / $30 output | GPT-5.6 Luna has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.6 LunaNot available | GPT-5.6 SolNot available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | GPT-5.6 LunaNot available | GPT-5.6 SolNot available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | GPT-5.6 Luna1M | GPT-5.6 Sol1M | Listed context windows are equal. |
Benchmark Deep Dive
AgenticGPT-5.6 Sol wins18 benchmarks
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 84.7% | 91.9% | GPT-5.6 Sol leads |
| BrowseCompSource | 83.3% | 92.2% | GPT-5.6 Sol leads |
| OSWorld 2.0Source | 45.6% | 62.6% | GPT-5.6 Sol leads |
| CyberGymSource | 77.9% | 84.5% | GPT-5.6 Sol leads |
| ExploitGymSource | 12.4% | 33.7% | GPT-5.6 Sol leads |
| ToolathlonSource | 53.4% | 58% | GPT-5.6 Sol leads |
| AA Agentic IndexSource | 45.6% | 54.0% | GPT-5.6 Sol leads |
| GDPval-AASource | 54.2% | 61.8% | GPT-5.6 Sol leads |
| GDPval-AASource | 1584 | 1736 | GPT-5.6 Sol leads |
| AA Harvey LABSource | 87.9% | 87.2% | GPT-5.6 Luna leads |
| AA ITBenchSource | 40.3% | 56.2% | GPT-5.6 Sol leads |
| AA Tau3 BankingSource | 27.2% | 33.0% | GPT-5.6 Sol leads |
| AA AutomationBenchSource | 42.2% | 51.2% | GPT-5.6 Sol leads |
| aaTerminalBench21Source | 80.9% | 88% | GPT-5.6 Sol leads |
| APEX-Agents-AASource | 35.8% | — | Not comparable |
| τ²-bench resultsSource | — | 85.1% | Not comparable |
| AA BriefcaseSource | — | 1501 | Not comparable |
| terminalBenchHardSource | — | 65.9% | Not comparable |
CodingGPT-5.6 Sol wins8 benchmarks
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol | Result |
|---|---|---|---|
| SWE-bench ProSource | 62.7% | 64.6% | GPT-5.6 Sol leads |
| Terminal-Bench 2.0Source | 84.7% | 91.9% | GPT-5.6 Sol leads |
| deepSweSource | 67.2% | 72.7% | GPT-5.6 Sol leads |
| FrontierCode 1.1 ExtendedSource | 55.1% | 60.6% | GPT-5.6 Sol leads |
| cursorBench32Source | 61.1% | 67.2% | GPT-5.6 Sol leads |
| AA Coding IndexSource | 71.5% | 77.4% | GPT-5.6 Sol leads |
| AA-SciCodeSource | 52.5% | 56.1% | GPT-5.6 Sol leads |
| VulcanBench v3Source | — | 87.0% | Not comparable |
Reasoning4 benchmarks
KnowledgeGPT-5.6 Sol wins10 benchmarks
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol | Result |
|---|---|---|---|
| GPQASource | 92.3% | 94.6% | GPT-5.6 Sol leads |
| GPQA-DSource | 92.3% | 94.6% | GPT-5.6 Sol leads |
| HealthBench ProfessionalSource | 55.7% | 60.5% | GPT-5.6 Sol leads |
| HealthBench HardSource | 32.0% | 33.1% | GPT-5.6 Sol leads |
| Artificial Analysis Intelligence IndexSource | 51.2% | 58.9% | GPT-5.6 Sol leads |
| AA-GPQA DiamondSource | 91.1% | 94.1% | GPT-5.6 Sol leads |
| AA-HLESource | 37.2% | 47.2% | GPT-5.6 Sol leads |
| AA-Omniscience IndexSource | -11.2% | 21.7% | GPT-5.6 Sol leads |
| AA-Omniscience AccuracySource | 41.5% | 58.5% | GPT-5.6 Sol leads |
| AA-Omniscience Hallucination RateSource | 90.1% | 88.8% | GPT-5.6 Sol leads |
MathGPT-5.6 Sol wins3 benchmarks
MultimodalGPT-5.6 Sol wins3 benchmarks
Inst. Following1 benchmarks
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol | Result |
|---|---|---|---|
| AA-IFBenchSource | — | 72.7% | Not comparable |
Frequently Asked Questions (6)
Which is better, GPT-5.6 Luna or GPT-5.6 Sol?
GPT-5.6 Luna and GPT-5.6 Sol are sibling variants in the GPT-5.6 family, so the right pick depends on whether you value the better benchmark line, cheaper tokens, or the larger context window. GPT-5.6 Sol is ahead on BenchLM's BenchAlign leaderboard 81.96 to 67.17.
Which is better for knowledge tasks, GPT-5.6 Luna or GPT-5.6 Sol?
GPT-5.6 Sol has the edge for knowledge tasks in this comparison, averaging 94.6 versus 92.3. Inside this category, AA-Omniscience Index is the benchmark that creates the most daylight between them.
Which is better for coding, GPT-5.6 Luna or GPT-5.6 Sol?
GPT-5.6 Sol has the edge for coding in this comparison, averaging 64.6 versus 62.7. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.
Which is better for math, GPT-5.6 Luna or GPT-5.6 Sol?
GPT-5.6 Sol has the edge for math in this comparison, averaging 87.5 versus 73.6. Inside this category, FrontierMath v2 (Tier 4) is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, GPT-5.6 Luna or GPT-5.6 Sol?
GPT-5.6 Sol has the edge for agentic tasks in this comparison, averaging 92 versus 84.1. Inside this category, GDPval-AA is the benchmark that creates the most daylight between them.
Which is better for multimodal and grounded tasks, GPT-5.6 Luna or GPT-5.6 Sol?
GPT-5.6 Sol has the edge for multimodal and grounded tasks in this comparison, averaging 83 versus 78.4. 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.