BenchLeader
Menu

GPT-5.6 Sol vs Step 5 Preview

Full profiles: GPT-5.6 Sol (OpenAI) and Step 5 Preview (StepFun): every benchmark, price and rank.

Updated · Data from 11 sources

As of 13 Nov 2026, GPT-5.6 Sol is ahead of Step 5 Preview on quality (65.3 vs 54.3 on the BenchLeader Index), but Step 5 Preview is 5.6× cheaper at $1.43 vs $8.00 per 1M tokens blended. Choose GPT-5.6 Sol for agents & tools (68 vs 50), coding (66 vs 50), and reasoning (65 vs 52); choose Step 5 Preview for fast first answers (22.8 s vs 68.5 s) and long context (69 vs 67), where it also streams faster.

Key facts (as of 13 Nov 2026)

  • GPT-5.6 Sol scores 65.3 on the BenchLeader Index and Step 5 Preview scores 54.3 (BenchLeader, 13 Nov 2026).
  • GPT-5.6 Sol costs $8.00 per 1M tokens blended (list price, 13 Nov 2026).
  • Step 5 Preview costs $1.43 per 1M tokens blended (list price, 13 Nov 2026).
  • Step 5 Preview streams faster than GPT-5.6 Sol (Artificial Analysis, 7-day median, 13 Nov 2026).
  • Of the 32 benchmarks both models have, GPT-5.6 Sol leads 28 (BenchLeader, 13 Nov 2026).
GPT-5.6 Sol benchmarks max OpenAI Step 5 Preview benchmarks StepFunAdd a third model

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingGPT-5.6 Sol66 vs 50 category score; IOI 91.2% vs 38.1%High
Agents & tool useGPT-5.6 Sol68 vs 50 category score; ITBench SRE (AA) 56.2% vs 55.6%High
MathsGPT-5.6 Sol70 vs 60 category score; ProofBench 83.0% vs 42.0%High
ReasoningGPT-5.6 Sol65 vs 52 category score; CritPt v2 hard subset (AA) 60.0% vs 32.5%High
KnowledgeGPT-5.6 Sol64 vs 54 category score; BioMysteryBench 71.1% vs 67.4%High
Long documentsStep 5 Preview69 vs 67 category score; AA-LCR v1.1 88.3% vs 84.0%Medium
Images and other mediaGPT-5.6 Sol66 vs 59 category score; MMMU-Pro 83.4% vs 76.4%Medium
Chat quality (human preference)–Only Step 5 Preview has an Arena rating (63 category score)Not comparable
Lowest costStep 5 Preview$1.43 vs $8.00 per 1M blended; agentic coding session $0.071 vs $0.320–
Fastest first answerStep 5 Preview22.8 s (Step 5 Preview) vs 68.5 s (GPT-5.6 Sol (max)) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • GPT-5.6 Sol (max) leads on quality: 65.3 vs 54.3.
  • GPT-5.6 Sol (max) is stronger in agents & tools, coding, knowledge, maths, multimodal, reasoning.
  • Step 5 Preview is stronger in long context.
  • Step 5 Preview is 5.6× cheaper ($1.43 vs $8.00 per 1M blended).
  • Step 5 Preview streams 1.4× faster (135 vs 94 tokens per second).

At a glance

SpecGPT-5.6 SolStep 5 Preview
DeveloperOpenAIStepFun
Released9 Jul 202618 Sep 2026
Context window (tokens)1,050,0001,000,000
Compared atGPT-5.6 Sol (max)Step 5 Preview
Open weightsNoYes
Price in / out / cached input per 1M$4.00 / $20.00 / $0.400$1.00 / $2.70 / $0.105

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryGPT-5.6 Sol (max)Step 5 PreviewLeader
Reasoning64.751.7GPT-5.6 Sol
Coding65.849.9GPT-5.6 Sol
Agents & tools67.649.7GPT-5.6 Sol
Maths70.060.1GPT-5.6 Sol
Knowledge63.853.7GPT-5.6 Sol
Instruction following71.2–not comparable
Human preference–63.1not comparable
Multimodal65.959.3GPT-5.6 Sol
Long context66.769.0Step 5 Preview
Composite–38.3not comparable

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75ReasoningCodingAgents & toolsMathsKnowledgeLong contextMultimodalGPT-5.6 Sol, Reasoning: 64.7GPT-5.6 Sol, Coding: 65.8GPT-5.6 Sol, Agents & tools: 67.6GPT-5.6 Sol, Maths: 70.0GPT-5.6 Sol, Knowledge: 63.8GPT-5.6 Sol, Long context: 66.7GPT-5.6 Sol, Multimodal: 65.9Step 5 Preview, Reasoning: 51.7Step 5 Preview, Coding: 49.9Step 5 Preview, Agents & tools: 49.7Step 5 Preview, Maths: 60.1Step 5 Preview, Knowledge: 53.7Step 5 Preview, Long context: 69.0Step 5 Preview, Multimodal: 59.3

Not drawn, measured for fewer than two of these models: Instruction following, Human preference.

Model, click to add or removeIndex

Where sources disagree

Across the 32 benchmarks both have, GPT-5.6 Sol leads 28, Step 5 Preview leads 4. Every overall index that rates both puts GPT-5.6 Sol ahead.

Overall indexGPT-5.6 Sol (max)Step 5 PreviewLeader
BenchLeader Index65.354.3GPT-5.6 Sol
Arena Text1485†1449GPT-5.6 Sol

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureGPT-5.6 Sol (max)Step 5 Preview
Blended price per 1M tokens$8.00$1.43
Input / output per 1M$4.00 / $20.00$1.00 / $2.70
Cached input per 1M$0.400$0.105
Price per task: chat reply$0.0076$0.0012
Price per task: summarize a 30-page report$0.060$0.014
Price per task: code edit$0.054$0.010
Price per task: agentic coding session$0.320$0.071
Price per task: structured extraction$0.012$0.0025
Agentic coding session, input mostly cached$0.158$0.030
Cost per benchmark task (AA Index)–$2.35
Output tokens per task (AA Index)–113k
Of which reasoning–80k
Time per task (AA Index)–14 min
Output speed94 tok/s135 tok/s
Time to first answer68.5 s22.8 s
Time to a complete answer157.3 s54.4 s

Each model is compared at its best-ranked reasoning-effort setting. GPT-5.6 Sol (max) scores 65.3 and takes 68 s to its first answer token; GPT-5.6 Sol (xhigh) scores 63.9 in 27 s, GPT-5.6 Sol (high) scores 63.4 in 16 s. Step 5 Preview scores 54.3 and takes 23 s to its first answer token. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

166 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherGPT-5.6 Sol (max)Step 5 Preview
BenchLeader IndexBenchLeader65.354.3
Agents & tools scoreBenchLeader67.649.7
Coding scoreBenchLeader65.849.9
Instruction following scoreBenchLeader71.2–
Knowledge scoreBenchLeader63.853.7
Long context scoreBenchLeader66.769.0
Maths scoreBenchLeader70.060.1
Multimodal scoreBenchLeader65.959.3
Reasoning scoreBenchLeader64.751.7
Composite scoreBenchLeader–38.3
Human preference scoreBenchLeader–63.1
Blended price $/MBenchLeader$8.00$1.43
Output speedBenchLeader94 tok/s135 tok/s
Time to first answerBenchLeader68.5 s22.8 s
Context windowBenchLeader1,050,000 tokens1,000,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub93.5%–
SimpleBenchSimpleBench64.8%†–
Arena Hard PromptsArena (LMArena)1511†1477
LiveBench 2.0 ReasoningLiveBench–83.8%
LiveBench 1.0 ReasoningLiveBench91.7%–
GPQA Diamond (AA)Artificial Analysis94.1%–
Humanity's Last Exam (AA)Artificial Analysis49.5%46.5%
HLE-Diamond (AA)Artificial Analysis–24.3%
GPQA Diamond (Vals)Vals AI95.2%–
EnigmaEvalScale AI SEAL37.1%†–
Kagi LLM BenchmarkKagi LLM Benchmark67.0%†–
ARC-AGI-1ARC Prize96.5%–
ARC-AGI-2ARC Prize92.5%–
ARC-AGI-3ARC Prize7.8%–
CritPtArtificial Analysis32.3%20.9%
CritPt v2 hard subset (AA)Artificial Analysis60.0%32.5%
MysteryMechanismVals AI33.3%18.5%
Chess PuzzlesEpoch AI Benchmarking Hub59.0%–
Mystery Game PuzzlesEpoch AI Benchmarking Hub58.0%–
LMCAEpoch AI Benchmarking Hub58.4%–
DTBenchEpoch AI Benchmarking Hub95.5%–
ForecastBenchForecastBench65.7%–
BTF-3Epoch AI Benchmarking Hub13.7%–
Coding
SciCodeSciCode57.1%58.9%
WeirdMLWeirdML87.0%–
Show all 166 benchmarks
MetricPublisherGPT-5.6 Sol (max)Step 5 Preview
Coding
FrontierCodeCognition47.5%†–
GSO-BenchGSO-Bench76.5%†–
Arena CodingArena (LMArena)1533†1505
Arena WebDevArena (LMArena)1615†1573†
LiveBench 2.0 CodingLiveBench–67.6%
LiveBench 1.0 CodingLiveBench83.9%–
LiveCodeBenchVals AI82.6%–
IOIVals AI91.2%38.1%
SWE-bench (Vals)Vals AI96.2%–
Code MigrationVals AI52.9%39.0%
Code Migration v1.1Vals AI–47.0%
ProgramBenchVals AI1.5%–
SWE-bench MultimodalVals AI–40.8%
Vibe Code Bench 1-100Vals AI20.0%–
Vibe Code Bench v1.1Vals AI80.5%70.4%
SWE Atlas: Codebase QnAScale AI SEAL46.0%†–
SWE Atlas: Test WritingScale AI SEAL45.9%†–
MirrorCodeEpoch AI Benchmarking Hub20.0%†–
Surface Evolver BenchEpoch AI Benchmarking Hub93.1%†–
DeepSWE v1.1Epoch AI Benchmarking Hub72.7%–
CursorBenchCursor41.7%–
ALE-BenchEpoch AI Benchmarking Hub2176.9–
FrontierSWEEpoch AI Benchmarking Hub32.2%–
Terminal-Bench 4.0 (AA)Artificial Analysis39.9%33.3%
Terminal-Bench 2.1 (AA)Artificial Analysis88.0%–
Startrise FrontendStartrise66.2%†–
GameBench 2 OverallDuelLab GameBench 266.7†–
GameBench 2 by reasoning settingDuelLab GameBench 279.7†–
Design Arena WebsiteDesign Arena1319†–
FrontierCode 1.1 (AA)Artificial Analysis–38.4%
FrontierCode 2.0 (AA)Artificial Analysis–12.0%
Agents & tools
Terminal-BenchTerminal-Bench38.5%–
OSWorld-Verified 2.0OSWorld27.3%–
APEX-AgentsMercor51.4%†–
Arena AgentArena (LMArena)4.6†-3.7
LiveBench Long AgenticLiveBench–19.4%
LiveBench BrowserLiveBench–24.8%
LiveBench Agentic CodingLiveBench56.2%–
Terminal-Bench HardArtificial Analysis65.9%–
τ²-Bench Telecom (AA)Artificial Analysis85.1%–
Terminal-Bench 2.1 (Vals)Vals AI85.8%–
HiL-BenchScale AI SEAL32.3%†–
τ³-bench Bankingτ²-bench46.9%†–
GDPval-AA v2.1, normalised scoreArtificial Analysis–54.3%
GDPval-AA v2.1Artificial Analysis–1592
τ³-Banking (AA)Artificial Analysis44.3%–
ITBench SRE (AA)Artificial Analysis56.2%55.6%
Analyst Agent (AA)Artificial Analysis47.5%35.0%
APEX-Agents (AA)Artificial Analysis–38.0%
CUA-benchVals AI8.3%–
CyberBenchVals AI76.3%–
Finance Agent v2Vals AI53.8%50.7%
Finance Agent v2.1Vals AI–58.9%
Harvey's Legal Agent Benchmark v1.1Vals AI–12.9%
Harvey's Legal Agent Benchmark v1Vals AI2.5%9.2%
Legal Research BenchVals AI48.1%40.4%
SkillsBenchVals AI54.1%–
SREBenchVals AI30.5%0.4%
Tax Agent BenchVals AI31.1%27.3%
Terminal-Bench 4.0 (Vals)Vals AI37.9%31.8%
Terminal-Bench ScienceVals AI20.0%2.9%
Time Horizon Index: KSPVals AI23.8%–
Web Search IndexVals AI43.6%–
Arena SearchArena (LMArena)1255†–
BALROGBALROG60.0%–
PostTrainBenchEpoch AI Benchmarking Hub31.0%–
GBAEvalEpoch AI Benchmarking Hub52.6%†–
Vending-Bench 2Andon Labs9619.4†–
AutomationBenchArtificial Analysis60.1%51.0%
GDP.pdf (AA)Artificial Analysis27.2%14.8%
EnterpriseOps-GymArtificial Analysis42.9%47.2%
AA-Briefcase v1.1Artificial Analysis14791423
Design Arena AgenticDesign Arena1231†–
Terminal-Bench-Science 0.1 (AA)Artificial Analysis22.4%2.4%
CWE-Bench (AA)Artificial Analysis–52.5%
DeepSec-Bench (AA)Artificial Analysis–15.8%
CyberGym end-to-end (AA)Artificial Analysis–74.8%
FACTS SearchGoogle92.9%†–
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub89.1%–
FrontierMath Tier 4Epoch AI Benchmarking Hub82.9%–
OTIS Mock AIMEEpoch AI Benchmarking Hub100.0%–
ProofBenchVals AI83.0%42.0%
LiveBench MathematicsLiveBench96.2%–
Arena MathsArena (LMArena)1494†1470
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub69.7%–
LiveBench 2.0 Data AnalysisLiveBench–63.5%
LiveBench ResearchLiveBench–75.5%
LiveBench 1.0 Data AnalysisLiveBench79.8%–
AA-Omniscience v1.1Artificial Analysis–16.4
MMLU-ProVals AI89.1%–
LegalBenchVals AI87.0%–
CorpFinVals AI64.4%–
TaxEvalVals AI74.8%–
PRBench FinanceScale AI SEAL50.5%–
PRBench LegalScale AI SEAL50.5%–
BioMysteryBenchVals AI71.1%67.4%
Excel Modeling BenchmarkVals AI72.3%57.1%
Excel Modeling Benchmark v1.1Vals AI–69.5%
MedCodeVals AI44.0%–
MedScribeVals AI85.2%–
Public Benefits Bench v1.1Vals AI66.5%58.1%
DrugDiscoveryBenchScale AI SEAL56.5%–
EBR-benchEpoch AI Benchmarking Hub44.8%–
GDP.pdfEpoch AI Benchmarking Hub30.7%–
AA-Omniscience v1.1: accuracyArtificial Analysis–36.4%
AA-Omniscience v1.1: non-hallucinationArtificial Analysis–68.7%
Instruction following
LiveBench LanguageLiveBench87.7%–
LiveBench Instruction FollowingLiveBench71.8%–
IFBenchArtificial Analysis72.7%–
Arena Instruction FollowingArena (LMArena)1490†1450
Human preference
Arena TextArena (LMArena)1485†1449
EQ-Bench 4EQ-Bench1236†–
Arena Creative WritingArena (LMArena)1471†1413
Arena Multi-turnArena (LMArena)1491†1445
Arena Longer QueriesArena (LMArena)1498†1461
Multimodal
Arena VisionArena (LMArena)1282†1263
LiveBench DesignLiveBench–70.8%
MMMU-ProArtificial Analysis83.4%76.4%
MMMU-Pro (Vals)Vals AI88.8%–
MortgageTaxVals AI67.3%–
GeoBenchGeoBench73.0%†–
Blueprint-Bench 2Epoch AI Benchmarking Hub33.6%†–
Code2Video-AAArtificial Analysis–821
AA-3DArtificial Analysis–859
Long context
AA-LCR v1.1Artificial Analysis84.0%88.3%
Arena DocumentArena (LMArena)1487†–
MLCRArtificial Analysis30.0%18.3%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub163.0†–
LiveBench 2.0LiveBench–54.9%
LiveBench 1.0LiveBench81.0%–
AA Intelligence Index v5.0Artificial Analysis–28.5
AA Intelligence Index v4.3.2Artificial Analysis47.043.7
AA Progress Index v1Artificial Analysis331.5318.2
Vals IndexVals AI–54.8
Vals Finance IndexVals AI–55.4%
Vals RSI IndexVals AI23.9%–
Vals Science IndexVals AI41.5%29.6%
Vals Multimodal IndexVals AI72.6%–
AA Cyber IndexArtificial Analysis–47.7
Safety & honesty
FORTRESSScale AI SEAL20.1%†–
Metadata
AA Openness IndexArtificial Analysis–33.3

Data as of 2026-11-13. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261110-2acac640a228 (https://benchmarks.duellab.org/releases/gb2-20261110-2acac640a228/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • GPT-5.6 Sol (max): 1 score revision.
  • Step 5 Preview: 1 score revision.
  • GPT-5.6 Sol (max): BenchLeader Index 65.8 → 65.3.
  • GPT-5.6 Sol (max): 1 score revision.
  • Step 5 Preview: 7 score revisions.
  • GPT-5.6 Sol (max): 1 score revision.
  • Step 5 Preview: 2 new results (Vals Index, Harvey's Legal Agent Benchmark v1.1).
  • Step 5 Preview: 1 score revision.
  • Step 5 Preview: BenchLeader Index 58.0 → 54.7.
  • GPT-5.6 Sol (max): 2 new results (AA Progress Index v1, CritPt v2 hard subset (AA)).

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

GPT-5.6 Sol vs Step 5 Preview: questions

Is GPT-5.6 Sol better than Step 5 Preview?
GPT-5.6 Sol (max) leads on quality: 65.3 vs 54.3. The BenchLeader Index combines every independent quality benchmark; GPT-5.6 Sol (max) is ahead overall as of 13 Nov 2026, but check the category scores for your use.
GPT-5.6 Sol vs Step 5 Preview: which is cheaper, and will that change?
Step 5 Preview is cheaper today: $1.43 against $8.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is GPT-5.6 Sol or Step 5 Preview better for coding?
GPT-5.6 Sol scores higher in coding (66 vs 50 on the category index, where 50 is average). On SciCode, the top shared coding benchmark, it is 57.1% for GPT-5.6 Sol against 58.9% for Step 5 Preview.
Which is faster, GPT-5.6 Sol or Step 5 Preview?
Step 5 Preview streams faster (135 vs 94 output tokens per second, Artificial Analysis 7-day median).
Is GPT-5.6 Sol or Step 5 Preview better for agents and tool use?
GPT-5.6 Sol scores higher in agentic tasks (68 vs 50 on the category index, where 50 is average).
More questions (4)
Which is better for long documents?
Step 5 Preview scores higher on long-context benchmarks (69 vs 67). Both take about 1.1M tokens: GPT-5.6 Sol 1,050,000, Step 5 Preview 1,000,000.
Which has lower latency?
Step 5 Preview answers first: 22.8 s against 68.5 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on GPT-5.6 Sol vs Step 5 Preview?
Across the 32 benchmarks both have, GPT-5.6 Sol leads 28, Step 5 Preview leads 4. Every overall index that rates both puts GPT-5.6 Sol ahead.
When was Step 5 Preview released?
Step 5 Preview was released on 18 Sep 2026.