Data as of November 13, 2026 · How the score is built
GPT-6.1 Sol
Decision readingGPT-6.1 Sol scores 84.6 out of 100 and ranks #5 of 215. This profile shows 55 source-displayable benchmark rows; its strongest eligible category is Knowledge at #3. API pricing is $2 input and $10 output per million tokens, with cached input at $0.1.
Email me when its price, rank or availability changes.
Capability
84.6/100
field median 49.4#5 of 215 ranked models
Public
#5of 215
Verified #4 of 81
Price
$2input / $10 output
input median $0.60cached $0.10 · blended $6
Speed
Not measured
field median 92 tok/sTime to first token not measured
Context
1.05Mtokens
field median 256,000Maximum output length is tracked separately
Holds badges in the November 2026 rankings release: #5 Overall, #5 Reasoning, #4 Knowledge (rank badges)
Decision snapshot
Where this model sits in the ranked field, its strongest published evidence, and what to validate before choosing it. The marker compares it with the current ranked catalog, not with an absolute quality threshold.
Strongest published evidence
Knowledge ranks #3. Particularly effective for knowledge-intensive tasks like research, analysis, and factual Q&A.
Validate before choosing
55 published rows leave some tracked benchmark slots empty. Independent runtime speed has not been measured.
Source-linked · 55 displayable benchmark rows
GPT-6.1 Sol API pricing
GPT-6.1 Sol costs $2.00 per million input tokens and $10.00 per million output tokens on the OpenAI API. Cached input is $0.10 per million.
| Token type | Price per 1M tokens |
|---|---|
| Input | $2.00 |
| Cached input | $0.10 |
| Output | $10.00 |
At these rates, a request with 10,000 input tokens and 1,000 output tokens costs $0.03 without caching.
On blended price (the average of the input and output rates), 1 of the 118 other ranked models with a published API price costs the same or less and scores higher. See where it sits on the chart
Pricing notes: OpenAI's GPT-6.1 Sol model page lists $2.00 input / $0.10 cached input / $2.50 cache writes / $10.00 output per million tokens. Prompts above 272K input tokens are charged at 2x input and cache rates and 1.5x output for the full request; regional processing adds 10% where available; Batch and Flex are 50% of Standard and Fast mode is 2x.
Rates from OpenAI GPT-6.1 Sol model documentation. API model ID: gpt-6.1-sol. Price registry updated November 12, 2026.
Is GPT-6.1 Sol right for your job?
Its place on the LLM Selector’s job rankings, the same order the job pages publish. Describe your own job to see where it stands for that.
Where GPT-6.1 Sol ranks by benchmark score:
We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy
Category score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #17 of 135Percentile 88thWeight 22%7 benchmarksVerified | 65.7 | 7 benchmarks | Verified | |||
| CodingRank #8 of 162Percentile 96thWeight 20%6 benchmarksVerified | 73.4 | 6 benchmarks | Verified | |||
| ReasoningRank #5 of 160Percentile 97thWeight 17%6 benchmarksVerified | 80.4 | 6 benchmarks | Verified | |||
| MultimodalRank #4 of 67Percentile 95thWeight 12%3 benchmarksVerified | 78.9 | 3 benchmarks | Verified | |||
| KnowledgeRank #3 of 180Percentile 99thWeight 12%5 benchmarksVerified | 85.6 | 5 benchmarks | Verified | |||
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MathRank Not rankedWeight 5%2 benchmarksVerified | 93.8 | 2 benchmarks | Verified |
55 of 917 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic7/7 verified
- Coding6/6 verified
- Reasoning6/6 verified
- Multimodal3/3 verified
- Knowledge5/5 verified
- MultilingualNot measured
- Inst. FollowingNot measured
- Math2/2 verified
Capability shape
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
GPT-6.1 Sol category percentile values
- Agentic88th percentile
- Coding96th percentile
- Reasoning97th percentile
- Multimodal95th percentile
- Knowledge99th percentile
- MultilingualNot eligible
- Instruction followingNot eligible
- MathNot eligible
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#17/135
- Coding#8/162
- Reasoning#5/160
- Multimodal#4/67
- Knowledge#3/180
- MultilingualNot ranked
- Inst. FollowingNot ranked
- MathNot ranked
Benchmark ledger
Every result says who ran the test. For this model, OpenAI ran 14 of 55. Coding opens by default. The tick marks the best source-verified result in the catalog; provisional leaders do not set the reference.
Coding6 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| DeepSWE | 71.9% | Best verified: Muse Wonder Full · 79.0% | Gap7.1 behind | Reported by OpenAI | Counts15% ref. weight |
| FrontierCode 2.0 | 39.0% | Best verified: Claude Fable 5.2 · 49.1% | Gap10.1 behind | Run by FrontierCode 2.0 | Counts4% ref. weight |
| VulcanBench Frontier v4 | 88.8% | Best verified: Grok 4.7 · 93.2% | Gap4.4 behind | Run by VulcanBench Frontier v4 | Counts3% ref. weight |
| ScarfBenchScarfBench: Cross-Framework Application Migration in Enterprise Java | 68.4% | Best verified: Gemini 4 Argon · 75.2% | Gap6.8 behind | Run by ScarfBench | Shown only |
| Bug Hunt Bench | 44.3 fixes | Best verified: Claude Fable 5.2 · 58.0 fixes | Gap13.7 behind | Run by Bug Hunt Bench | Shown only |
| PostTrainBenchPostTrainBench v1.2 | 38.8% | Best verified: Claude Opus 5.5 · 43.8% | Gap5 behind | Run by PostTrainBench | Shown only |
Agentic7 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| Terminal-Bench 4.1 | 60.90% | Best verified: Claude Opus 5.5 · 65.80% | Gap4.9 behind | Run by Terminal-Bench 4.1 | Counts9% ref. weight |
| AutomationBench | 36.1% | Best verified: MiniMax M3.1 Flash · 62.2% | Gap26.1 behind | Reported by OpenAI | Counts6% ref. weight |
| Terminal-Bench-Science 0.1 | 57.0% | Best verified: Claude Mythos 5.2 · 75.3% | Gap18.3 behind | Reported by OpenAI | Shown only |
| ExploitGym | 35.1% | Best verified: Gemini 4 Carbon · 53.5% | Gap18.4 behind | Reported by OpenAI | Shown only |
| MazeBenchMazeBench: visual spatial reasoning in a 3D open world | 17% | Best verified: GPT-6 Astra · 23% | Gap6 behind | Run by MazeBench | Shown only |
| Hermes IndexHermes Index: agentic capability and cost per task inside Hermes Agent | 57.1% | Best verified: Claude Opus 5.5 · 63.3% | Gap6.2 behind | Run by Hermes Index | Shown only |
| Hermes BenchHermes Bench: 150 agent tasks graded on the files and state the agent leaves behind | 73.9% | Best verified: Claude Opus 5.5 · 76.7% | Gap2.8 behind | Run by Hermes Bench | Shown only |
Reasoning6 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| ARC-AGI-2Abstraction and Reasoning Corpus for AGI v2 | 94.2% | Best verified: GPT-6 Astra · 95% | Gap0.8 behind | Run by ARC-AGI-2 | Counts15% ref. weight |
| ARC-AGI-3Abstraction and Reasoning Corpus for AGI v3 | 52.7% | Best verified: Claude Fable 5.2 · 65.7% | Gap12.9 behind | Run by ARC-AGI-3 | Counts15% ref. weight |
| BioBasicsBenchBioBasicsBench: everyday wet-lab science | 62.6% | Best verified: GPT-6 Astra · 67.0% | Gap4.4 behind | Reported by OpenAI | Shown only |
| STEMBenchSTEMBench120: six-domain scientific research progress | 2.76 | Best verified: GPT-6 Astra · 2.81 | Gap0.1 behind | Run by STEMBench | Shown only |
| ARC-AGI-1ARC-AGI-1 Semi-Private Evaluation | 96.50% | Best verified: Claude Fable 5 · 98.50% | Gap2 behind | Run by ARC-AGI-1 | Shown only |
| ARC-AGI-3 Provider AdapterARC-AGI-3 Semi-Private: Provider Adapter harness | 96.18% | Best verified: GPT-6 Astra · 98.55% | Gap2.4 behind | Run by ARC-AGI-3 Provider Adapter | Shown only |
Multimodal3 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| ParseBenchParseBench: document parsing for AI agents (LlamaIndex leaderboard) | 69.63% | Best verified: Claude Opus 5.5 · 78.80% | Gap9.2 behind | Run by ParseBench | Counts7% ref. weight |
| ExtractBench (LlamaIndex)ExtractBench: schema-guided enterprise document extraction (LlamaIndex leaderboard) | 95.31% | Best verified: GPT-6.1 Sol · 95.31% | GapBest verified | Run by ExtractBench (LlamaIndex) | Shown only |
| ExtractBench ProExtractBench Pro: long-tail document extraction challenges (LlamaIndex) | 12% | Best verified: GPT-6 Astra · 58% | Gap46 behind | Run by ExtractBench Pro | Shown only |
Knowledge5 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| HealthBench (raw)HealthBench raw score | 56.7% | Best verified: Claude Sonnet 5.5 · 69.4% | Gap12.7 behind | Reported by OpenAI | Shown only |
| HealthBench (length-adjusted)HealthBench length-adjusted score | 58.5% | Best verified: Claude Sonnet 5.5 · 65.4% | Gap6.9 behind | Reported by OpenAI | Shown only |
| HealthBench Professional | 64.2% | Best verified: Claude Sonnet 5.5 · 69.2% | Gap5 behind | Reported by OpenAI | Shown only |
| HealthBench Professional (raw)HealthBench Professional raw score | 67.2% | Best verified: Claude Sonnet 5.5 · 77.1% | Gap9.9 behind | Reported by OpenAI | Shown only |
| HealthBench Hard | 36.2% | Best verified: Muse Spark · 42.8% | Gap6.6 behind | Reported by OpenAI | Shown only |
Math2 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| FrontierMath v2 (Tiers 1-3)FrontierMath v2 Tiers 1-3 | 93.684% | Best verified: Claude Fable 5.2 · 95.088% | Gap1.4 behind | Run by FrontierMath v2 (Tiers 1-3) | CountsWeighted 30% |
| FrontierMath v2 (Tier 4)FrontierMath v2 Tier 4 | 100.000% | Best verified: Claude Fable 5.2 · 100.000% | GapBest verified | Run by FrontierMath v2 (Tier 4) | CountsWeighted 10% |
External signals26 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| Game Arena Go | 1448 | Best verified: Claude Fable 5.2 · 1877 | Gap429 behind | Run by Game Arena Go | Shown only |
| ASCII Arena | 1686 | Best verified: Gemini 4 Carbon · 2019 | Gap333 behind | Run by ASCII Arena | Shown only |
| Landmark Arena | 1694 | Best verified: Claude Opus 5.5 · 1792 | Gap98 behind | Run by Landmark Arena | Shown only |
| Pelican Arena | 1787 | Best verified: GPT-6 Astra · 1819 | Gap32 behind | Run by Pelican Arena | Shown only |
| FrontierPhysicsEnd-to-End Physics Research Benchmark for AI Agents | 19.6% | Best verified: Claude Opus 5.5 · 27.4% | Gap7.8 behind | Run by FrontierPhysics | Shown only |
| Cookie-BenchCookie-Bench: Continuous On-screen Key Interaction Evaluation for Web Generation | 89.7% | Best verified: Claude Opus 5.5 · 91.5% | Gap1.8 behind | Lab-reported, from another publisher’s document | Shown only |
| BlenderWorld (text-only)BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References (text-only subset) | 85.3% | Best verified: Claude Opus 5.5 · 86.1% | Gap0.8 behind | Lab-reported, from another publisher’s document | Shown only |
| NYT ConnectionsNYT Connections LLM Benchmark (extended version) | 97.8% | Best verified: Gemini 4 Carbon · 99.3% | Gap1.5 behind | Run by NYT Connections | Shown only |
| Joke Arena | 1519 | Best verified: GPT-6 Astra · 1564 | Gap45 behind | Run by Joke Arena | Shown only |
| Chart Arena | 1627 | Best verified: GPT-6.1 Sol · 1627 | GapBest verified | Run by Chart Arena | Shown only |
| Landing Page Arena | 1691 | Best verified: GPT-6.1 Sol · 1691 | GapBest verified | Run by Landing Page Arena | Shown only |
| DAYJOB Healthcare | 10.8% | Best verified: Claude Opus 5.5 · 25.2% | Gap14.4 behind | Run by DAYJOB Healthcare | Shown only |
| DAYJOB | 15.5% | Best verified: Claude Opus 5.5 · 24.5% | Gap9 behind | Run by DAYJOB | Shown only |
| FACTS Multimodal | 52.6% | Best verified: Gemini 4 Carbon · 60.2% | Gap7.6 behind | Run by FACTS Multimodal | Shown only |
| StealthMarkStealthMark Intelligence Benchmarks (overall score) | 67.8 | Best verified: GPT-6 Astra · 74.6 | Gap6.8 behind | Run by StealthMark | Shown only |
| Clocktower BenchClocktower Bench (Blood on the Clocktower) | 1734 | Best verified: GPT-6 Astra · 1859 | Gap125 behind | Run by Clocktower Bench | Shown only |
| ExploitBenchExploitBench v8-bench | 100% | Best verified: GPT-6 Astra · 100% | Gap0.3 behind | Reported by OpenAI | Shown only |
| SEC-Bench Pro | 78.8% | Best verified: Gemini 4 Carbon · 88.2% | Gap9.4 behind | Reported by OpenAI | Shown only |
| ExploitBench Internal PortExploitBench — Internal Port (June–August 2026) | 21.5% | Best verified: GPT-6 Cyber · 42.0% | Gap20.5 behind | Reported by OpenAI | Shown only |
| ExploitGym HoneypotExploitGym Honeypot out-of-scope exploitation rate | 0.00% | Best verified: GPT-6 Astra · 0.00% | GapHighest listed | Reported by OpenAI | Shown only |
| SpireClimbSpireClimb (Slay the Spire 1 and 2) | 23.8% | Best verified: GPT-6 Astra · 34.0% | Gap10.2 behind | Run by SpireClimb | Shown only |
| Chess PuzzlesChess Puzzles (Epoch AI) | 68.0% | Best verified: Claude Fable 5.2 · 79.0% | Gap11 behind | Run by Chess Puzzles | Shown only |
| Mystery Game PuzzlesMystery Game Puzzles (Epoch AI) | 80.0% | Best verified: Claude Fable 5.2 · 90.0% | Gap10 behind | Run by Mystery Game Puzzles | Shown only |
| Furniture AssemblyFurniture Assembly Benchmark (Epoch AI) | 80.0% | Best verified: Claude Fable 5.2 · 90.0% | Gap10 behind | Run by Furniture Assembly | Shown only |
| EBR-benchEarthborne Rangers (EBR-bench) | 54.3% | Best verified: Claude Fable 5.2 · 90.5% | Gap36.2 behind | Run by EBR-bench | Shown only |
| MirrorCodeMirrorCode (Epoch AI and METR) | 61.1% | Best verified: Claude Sonnet 5.5 · 97.8% | Gap36.7 behind | Run by MirrorCode | Shown only |
Outlined bars are lab-reported. Filled bars are benchmark-run or independent.
What it costs to get this score
Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.
GPT-6.1 Sol · 84.6 score · $6 blended per million tokens
The chart opens on the current model. Scroll horizontally to inspect the full price axis.
Horizontal: blended price per million tokens, log scale · Vertical: public score
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- gpt-6.1-solOpenAI GPT-6.1 Sol model documentation
- Context window
- 1.05MOpenAI GPT-6.1 Sol model documentation
- Maximum output
- 128,000 tokensOpenAI GPT-6.1 Sol model documentation
- Knowledge cutoff
- April 30, 2026OpenAI GPT-6.1 Sol model documentation
- Input modalities
- text, imageOpenAI GPT-6.1 Sol model documentation
- Output modalities
- textOpenAI GPT-6.1 Sol model documentation
- Parameters
- Not disclosed by the provider
- Availability
- OpenAI Responses API · OpenAI Chat Completions APIOpenAI GPT-6.1 Sol model documentation
Use the Responses API for tool calling. Chat Completions supports GPT-6.1 Sol without tool calling. OpenAI GPT-6.1 Sol model documentation, read Sep 29, 2026
- Cloud regions
- Not tracked yet
- Lifecycle
- activeOpenAI GPT-6.1 Sol model documentation
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Published at $0.10 per million cached input tokensOpenAI GPT-6.1 Sol model documentation
- Self-host
- Weights are not published
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
GPT-6.1 Sol ranks #5 of 215 on the public leaderboard with a score of 84.63/100. Its source-verified position is #4 of 81.
GPT-6.1 Sol is a proprietary model with a 1.05M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
OpenAI documents GPT-6.1 Sol as an API model. It supports reasoning effort from low to max; tool calling is supported through the Responses API, while Chat Completions is supported without tool calling.
OpenAI's September 29, 2026 launch charts report exact DeepSWE v1.1, AutomationBench 1.0.6, and Terminal-Bench Science 0.1 results. We use Max effort for the canonical row, even when a lower effort scores higher. The system card adds separate raw and length-adjusted HealthBench results, ExploitGym, ExploitBench, and SEC-Bench Pro. OpenAI also reports 71.4% partial reward on the OSWorld 2.0 offline set, which we keep out of the binary-completion field. GDP.pdf, HealthBench Consensus, MentalHealthBench, and internal safety evaluations remain in the cited sources because their protocols have no confirmed matching field. These are provider-run results; the cyber scores are display-only.
GPT-6.1 Sol sits in the GPT-6 family with GPT-6 Astra, GPT-6.1 Terra, GPT-6 Sol, GPT-6 Luna, GPT-6 Cyber. Its explicit predecessor is GPT-6 Sol. 55 of 917 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Knowledge at #3, while its lowest eligible position is Agentic at #17. particularly effective for knowledge-intensive tasks like research, analysis, and factual Q&A.
Last updated November 13, 2026. Runtime fields remain blank until a sourced snapshot exists.
Deployment options
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Where to run GPT-6.1 Sol by API
OpenRouter is the one host we track that publishes a per-token price for it. Prices read from each host’s own pages on November 11, 2026.
| Host | Input /1M | Output /1M | Cached /1M | Context |
|---|---|---|---|---|
| OpenRouter | $2.00 | $10.00 | $0.10 | 1.05M |
Questions
How does GPT-6.1 Sol perform overall in AI benchmarks?
GPT-6.1 Sol ranks #5 out of 215 models on the public BenchAlign leaderboard, with a score of 84.63/100. Its evidence status is Supported, and this profile shows 55 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is GPT-6.1 Sol good for knowledge and understanding?
GPT-6.1 Sol ranks #3 out of 180 eligible models for knowledge and understanding, with a public category score of 85.6/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-6.1 Sol good for coding and programming?
GPT-6.1 Sol ranks #8 out of 162 eligible models for coding and programming, with a public category score of 73.4/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-6.1 Sol good for mathematics?
GPT-6.1 Sol has source-displayable benchmark coverage for mathematics, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is GPT-6.1 Sol good for reasoning and logic?
GPT-6.1 Sol ranks #5 out of 160 eligible models for reasoning and logic, with a public category score of 80.4/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-6.1 Sol good for agentic tool use and computer tasks?
GPT-6.1 Sol ranks #17 out of 135 eligible models for agentic tool use and computer tasks, with a public category score of 65.7/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-6.1 Sol good for multimodal and grounded tasks?
GPT-6.1 Sol ranks #4 out of 67 eligible models for multimodal and grounded tasks, with a public category score of 78.9/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.
Which sibling models are related to GPT-6.1 Sol?
GPT-6.1 Sol belongs to the GPT-6 family. Related tracked variants include GPT-6 Astra, GPT-6.1 Terra, GPT-6 Sol, plus 2 more. A sibling link indicates shared lineage or a documented configuration relationship; it does not mean the variants have identical pricing, context limits, benchmark evidence, or deployment behavior. Compare before switching.
Does GPT-6.1 Sol have full benchmark coverage on BenchLM?
No. GPT-6.1 Sol currently has 75 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of GPT-6.1 Sol?
GPT-6.1 Sol has a documented context window of 1.05M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.