GPT-6.1 Sol vs Step 3.7 Flash
Updated October 21, 2026. 11 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
vs boards won of 11 shared
Decision reading
Only GPT-6.1 Sol has a coding score, 82.4 on the Vibe Coding Index; Step 3.7 Flash has none, so the two cannot be ranked on coding.
Step 3.7 Flash has the edge where output tokens cost 8.7x less.
Model AGPT-6.1 SolOpenAI82.4/1005 independent benchmarksNo cost per solved taskModel BStep 3.7 FlashStepFunNot benchmarkedNo coding benchmarkNo cost per solved task
- Shared results
- 11
- GPT-6.1 Sol only
- 84
- Step 3.7 Flash only
- 7
- Like-for-like categories
- 5 of 9
Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Only GPT-6.1 Sol has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
GPT-6.1 Sol
GPT-6.1 Sol leads, 51.8 against 19.5.
Same board, both measured
Web design
Design Arena: websites
GPT-6.1 Sol
GPT-6.1 Sol leads, 1320 against 1196.
Same board, both measured
General
Vals Index
No call
Only GPT-6.1 Sol has a Vals Index result, 63.9%.
Directional only
Long documents
Listed context window
GPT-6.1 Sol
GPT-6.1 Sol takes 1.05M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Step 3.7 Flash
GPT-6.1 Sol $0.0070, Step 3.7 Flash $0.0008 per turn.
List prices
At a glance
- Coding index
- GPT-6.1 Sol 82.4 from 5 benchmarks, Step 3.7 Flash not benchmarked.
- Widest category
- Coding, 40.1 points apart on average over 2 shared benchmarks.
- Widest single gap
- ALE-Bench: GPT-6.1 Sol 2743.8 against Step 3.7 Flash 694.1, 68.3 points apart on a 0 to 100 scale.
- Output price
- GPT-6.1 Sol $10.00 against Step 3.7 Flash $1.15 per million output tokens, 8.7x apart.
- Context window
- GPT-6.1 Sol takes 1.05M tokens, Step 3.7 Flash 262K.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | GPT-6.1 Sol | Bars | Step 3.7 Flash | Ties |
|---|---|---|---|---|
| Coding | 2 | 0 | ||
| Intelligence | 1 | 0 | ||
| Reasoning | 1 | 0 | ||
| Web design | 6 | 0 | ||
| General | 1 | 0 | ||
| All shared boards | 11 | 0 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index20
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| LiveBench coding | 96.0% | Not measured | Only GPT-6.1 Sol measured |
| LMArena WebDev | 1753 | Not measured | Only GPT-6.1 Sol measured |
| AA Coding Index | Not measured | 39.6 | Only Step 3.7 Flash measured |
| Terminal-Bench 4 (Vals AI) | 55.0% | Not measured | Only GPT-6.1 Sol measured |
| Vibe Code Bench (Vals AI) | 88.9% | Not measured | Only GPT-6.1 Sol measured |
| Code migration (Vals AI) | 65.1% | Not measured | Only GPT-6.1 Sol measured |
| ProgramBench (Vals AI) | 3.0% | Not measured | Only GPT-6.1 Sol measured |
| IOI (Vals AI) | 96.9% | Not measured | Only GPT-6.1 Sol measured |
| SWE-bench Multimodal (Vals AI) | 43.5% | Not measured | Only GPT-6.1 Sol measured |
| CyberBench (Vals AI) | 39.3% | Not measured | Only GPT-6.1 Sol measured |
| SciCode | 55.8% | 43.9% | GPT-6.1 Sol |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 35.6% | Only Step 3.7 Flash measured |
| FrontierCode | 50.2% | Not measured | Only GPT-6.1 Sol measured |
| ALE-Bench | 2743.8 | 694.1 | GPT-6.1 Sol |
| LMArena Text: coding | 1478 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: software and IT | 1476 | Not measured | Only GPT-6.1 Sol measured |
| LMArena WebDev: React | 1754 | Not measured | Only GPT-6.1 Sol measured |
| LMArena WebDev: HTML | 1800 | Not measured | Only GPT-6.1 Sol measured |
| LMArena WebDev: from an image | 1702 | Not measured | Only GPT-6.1 Sol measured |
| Neither model has a result on the other 19 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| AA Intelligence Index | 51.8 | 19.5 | GPT-6.1 Sol |
| Epoch Capabilities Index | 168.4 | Not measured | Only GPT-6.1 Sol measured |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index18 boards, none shared
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| Design Arena: agents building mobile apps | 1200 | Not measured | Only GPT-6.1 Sol measured |
| Design Arena: spreadsheets | 1244 | Not measured | Only GPT-6.1 Sol measured |
| LiveBench research | 83.5% | Not measured | Only GPT-6.1 Sol measured |
| LiveBench browser | 88.0% | Not measured | Only GPT-6.1 Sol measured |
| LiveBench long agentic | 73.0% | Not measured | Only GPT-6.1 Sol measured |
| SREBench (Vals AI) | 50.8% | Not measured | Only GPT-6.1 Sol measured |
| Finance Agent (Vals AI) | 52.0% | Not measured | Only GPT-6.1 Sol measured |
| Tax Agent Bench (Vals AI) | 21.8% | Not measured | Only GPT-6.1 Sol measured |
| Legal Research Bench (Vals AI) | 38.5% | Not measured | Only GPT-6.1 Sol measured |
| Harvey's Legal Agent Benchmark (Vals AI) | 6.3% | Not measured | Only GPT-6.1 Sol measured |
| Excel Modeling Benchmark (Vals AI) | 70.8% | Not measured | Only GPT-6.1 Sol measured |
| BioMysteryBench (Vals AI) | 79.6% | Not measured | Only GPT-6.1 Sol measured |
| Terminal-Bench-Science (Vals AI) | 52.9% | Not measured | Only GPT-6.1 Sol measured |
| Vals Science Index | 59.6% | Not measured | Only GPT-6.1 Sol measured |
| Robo-bench (Vals AI) | 28.3% | Not measured | Only GPT-6.1 Sol measured |
| τ²-Bench Telecom | Not measured | 98.5% | Only Step 3.7 Flash measured |
| APEX-Agents | 60.0% | Not measured | Only GPT-6.1 Sol measured |
| SpireClimb | 24.7% | Not measured | Only GPT-6.1 Sol measured |
| Neither model has a result on the other 21 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| LiveBench reasoning | 92.6% | Not measured | Only GPT-6.1 Sol measured |
| MysteryMechanism (Vals AI) | 46.4% | Not measured | Only GPT-6.1 Sol measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 80.9% | Only Step 3.7 Flash measured |
| Humanity's Last Exam (Artificial Analysis run) | 52.9% | 21.4% | GPT-6.1 Sol |
| GPQA Diamond (Epoch AI run) | 95.4% | Not measured | Only GPT-6.1 Sol measured |
| ARC-AGI | 98.5% | Not measured | Only GPT-6.1 Sol measured |
| ARC-AGI-2 | 94.2% | Not measured | Only GPT-6.1 Sol measured |
| Chess puzzles | 68.0% | Not measured | Only GPT-6.1 Sol measured |
| Mystery game puzzles | 80.0% | Not measured | Only GPT-6.1 Sol measured |
| EBR-bench | 54.3% | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: hard prompts | 1463 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: expert | 1489 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: hard prompts in English | 1465 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Vision: diagrams | 1306 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Vision: homework | 1356 | Not measured | Only GPT-6.1 Sol measured |
| Neither model has a result on the other 6 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| OTIS Mock AIME | 100.0% | Not measured | Only GPT-6.1 Sol measured |
| FrontierMath, tiers 1 to 3 | 93.7% | Not measured | Only GPT-6.1 Sol measured |
| FrontierMath, tier 4 | 100.0% | Not measured | Only GPT-6.1 Sol measured |
| ProofBench | 99.0% | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: math | 1489 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: mathematical work | 1513 | Not measured | Only GPT-6.1 Sol measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| MedCode (Vals AI) | 48.8% | Not measured | Only GPT-6.1 Sol measured |
| SimpleQA Verified | 73.9% | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: science | 1444 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: medicine and healthcare | 1433 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: legal and government | 1471 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: business and finance | 1445 | Not measured | Only GPT-6.1 Sol measured |
| Neither model has a result on the other 11 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index60
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| Design Arena: websites | 1320 | 1196 | GPT-6.1 Sol |
| Design Arena: UI components | 1330 | 1179 | GPT-6.1 Sol |
| Design Arena: game dev | 1389 | 1158 | GPT-6.1 Sol |
| Design Arena: data viz | 1280 | 1172 | GPT-6.1 Sol |
| Design Arena: 3D | 1463 | 1129 | GPT-6.1 Sol |
| Design Arena: all code categories | 1345 | 1182 | GPT-6.1 Sol |
| Design Arena: SVG | Not measured | 1085 | Only Step 3.7 Flash measured |
| Design Arena: ASCII art | Not measured | 1152 | Only Step 3.7 Flash measured |
| LiveBench design | 82.1% | Not measured | Only GPT-6.1 Sol measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| Vals Index | 63.9% | Not measured | Only GPT-6.1 Sol measured |
| LiveBench data analysis | 82.7% | Not measured | Only GPT-6.1 Sol measured |
| Public Benefits Bench (Vals AI) | 59.3% | Not measured | Only GPT-6.1 Sol measured |
| SAGE (Vals AI) | 46.5% | Not measured | Only GPT-6.1 Sol measured |
| MedScribe (Vals AI) | 86.5% | Not measured | Only GPT-6.1 Sol measured |
| IFBench | Not measured | 67.3% | Only Step 3.7 Flash measured |
| Long context reasoning (AA-LCR) | 84.0% | 73.7% | GPT-6.1 Sol |
| DTBench | 96.5% | Not measured | Only GPT-6.1 Sol measured |
| LMCA | 62.8% | Not measured | Only GPT-6.1 Sol measured |
| Furniture assembly | 80.0% | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text | 1443 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: instruction following | 1466 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: creative writing | 1425 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Vision | 1288 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: English | 1450 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: non-English | 1433 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: Chinese | 1487 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: French | 1460 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: Spanish | 1439 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: Russian | 1457 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: multi-turn | 1442 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: longer queries | 1458 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: writing and language | 1444 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Text: entertainment and media | 1426 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Vision: English | 1274 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Vision: reading text | 1289 | Not measured | Only GPT-6.1 Sol measured |
| LMArena Vision: creative writing | 1301 | Not measured | Only GPT-6.1 Sol measured |
| Neither model has a result on the other 19 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-6.1 Sol | Step 3.7 Flash | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
GPT-6.1 Sol$0.0070Fits in one request
Step 3.7 Flash$0.0008Fits in one request
Step 3.7 Flash costs less
Repository review
50K input + 3K output tokens
GPT-6.1 Sol$0.130Fits in one request
Step 3.7 Flash$0.013Fits in one request
Step 3.7 Flash costs less
Agent loop
200K input + 10K output tokens
GPT-6.1 Sol$0.500Fits in one request
Step 3.7 Flash$0.051Fits in one request
Step 3.7 Flash costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | OpenAI GPT-6.1 Sol | StepFun Step 3.7 Flash |
|---|---|---|
| Maker | OpenAI | StepFun |
| Weights | No public weights listed | Open weights |
| Released | September 29, 2026 | May 28, 2026 |
| Context window | 1.05M | 262K |
| Input $/Mtok | $2.00 | $0.20 |
| Output $/Mtok | $10.00 | $1.15 |
| Coding index | 82.4 | Not benchmarked |
| Independent measurements behind it | 5 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-6.1 Sol better than Step 3.7 Flash for coding?
Only GPT-6.1 Sol has a coding score, 82.4 on the Vibe Coding Index; Step 3.7 Flash has none, so the two cannot be ranked on coding. Step 3.7 Flash has the edge where output tokens cost 8.7x less.
Which is cheaper, GPT-6.1 Sol or Step 3.7 Flash?
GPT-6.1 Sol costs $2.00 per million input tokens and $10.00 per million output tokens; Step 3.7 Flash costs $0.20 and $1.15. Cost per solved task is known for neither of them.
Which has the bigger context window, GPT-6.1 Sol or Step 3.7 Flash?
GPT-6.1 Sol, with 1.05M tokens against 262K.
How many benchmarks is the GPT-6.1 Sol vs Step 3.7 Flash comparison based on?
11 benchmarks have a result for both models as of October 21, 2026. GPT-6.1 Sol's coding index rests on 5 independent benchmarks and Step 3.7 Flash's on 0. Where only one model has a result, the page says so instead of naming a winner.