GPT-4o
Language modelActive
GPT-4o
OpenAI
- Released
- May 13, 2024
- Data date
- September 8, 2026
- Successor
- GPT-4.1
- Speed
- 154 tokens/sOutput throughput
- Price
- Input · Output / 1M tokens
- Input
- Text, Image
- Output
- Text
GPT-4o brought text, image, and audio processing into one model. Earlier voice interactions passed through separate models for transcription, the text response, and speech generation. The unified model could also account for tone and other information in the audio.
That multimodal design went beyond the text and image inputs of the API version covered here. GPT-4.1 later emphasized coding, instruction following, and a much longer context window.
Specifications and access
| Specification | Value and source |
|---|---|
| Weights | ProprietarySource |
| Access | APISource |
| Parameters | 200B (estimated)Source |
| Context window | 128,000 tokensSource |
| Knowledge cutoff | October 2023Source |
| Notes | EpochAI estimates ~200B parametersSource |
| API model ID | gpt-4oSource |
| Maximum output | 16,384 tokensSource |
| Knowledge cutoff | 2023-10-01Source |
| Input | Text, ImageSource |
| Output | TextSource |
| Reasoning mode | noneSource |
| API interfaces | Responses API, Chat Completions APISource |
| Tool calling | YesSource |
| Structured output | YesSource |
| Release | 2024-05-13Source |
Page 1 of 14
Published benchmarks
Reasoning
LongBench v2
503 tasks with contexts from 8,000 to 2 million words cover documents, dialogue, code repositories, and structured data. Scores with chain of thought are not interchangeable with direct-answer results without reasoning.
| Sub-task | GPT-4o (This model) | Gemini 2.5 Pro | Gemini 2.5 Flash | o1 | DeepSeek-R1 |
|---|---|---|---|---|---|
| Easy | 54.2% | 75% (best value in this row) | 72.3% | 66.8% | 66.1% |
| Hard | 49.7% | 56.1% (best value in this row) | 55.8% | 52.1% | 53.4% |
| Long, 128,000 to 2 million words | 43.5% | 71% (best value in this row) | 55.7% | 58.1% | 59.3% |
| Medium, 32,000 to 128,000 words | 48.6% | 56.3% | 60% (best value in this row) | 53.5% | 54.4% |
| Overall | 51.4% | 63.3% (best value in this row) | 62.1% | 57.7% | 58.3% |
| Short, up to 32,000 words | 59.6% | 67.2% | 68.3% (best value in this row) | 62.6% | 62.2% |
higher is betterAs of September 2, 2026Source: LongBenchbars scaled within each row, not across rows
Methodology and context
- Metric
- accuracy with chain of thought
GPQA Diamond
| Sub-task | GPT-4o (This model) | GPT-4o mini |
|---|---|---|
| Few-Shot CoT | 54.04% (best value in this row) | 44.95% |
| Zero-Shot CoT | 53.54% (best value in this row) | 43.43% |
higher is betterAs of August 11, 2026Source: Vals AIbars scaled within each row, not across rows
Methodology and context
- Metric
- accuracy
TutorBench
Overall
1,490 tasks test adaptive explanations, feedback, and hints across six STEM subjects. A Claude 4 Sonnet judge scores weighted rubrics, so strong results do not prove safe tutoring of real learners.
higher is betterAs of August 26, 2026Scale: 0% - 100%
Methodology and context
- Comparison cohort
- scale-tutorbench-overall
- Metric
- weighted rubric score
- Scale
- 0-100
- Participants
- 8
- Methodology and context
- 1,490 tasks with weighted rubrics and Claude 4 Sonnet as judge.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Evaluator: Scale AI. Dataset type: public. Attempts: pass@1. Score display: 36.12% ± 0.96. Source ID: scale-ai. Evidence capture ID: curated-scale-ai-2026-08-26-2d47f3f45c11. Source payload hash: 2d47f3f45c11d5dfb09f242e535afc28a9fd65f5672d1a6a41964d0c2ec8c719. Parser version: curated-v1. Imported at: 2026-09-29
Coding
LiveCodeBench v6
| Sub-task | GPT-4o (This model) | GPT-4o mini |
|---|---|---|
| Easy | 86.03% (best value in this row) | 60.87% |
| Hard | 5.14% (best value in this row) | 1.43% |
| Medium | 39.16% (best value in this row) | 16.97% |
higher is betterAs of August 11, 2026Source: Vals AIbars scaled within each row, not across rows
Methodology and context
- Metric
- accuracy
Recorded measurement details
LongBench v2: Easy (Show measurement, test conditions, and source)
- Source value
- 54.2
- Score
- 54.2
- Metric
- accuracy with chain of thought
- Unit
- %
- Benchmark version
- v2
- Category
- reasoning
- Task
- easy
- Task label
- Easy
- Direction
- Higher is better
- Comparison cohort
- longbench-v2-cot-easy
- Source type
- official-leaderboard
- Evaluator
- LongBench
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: LongBench. Dataset type: public. Leaderboard date: 2024-11-20. Attempts: pass@1. Reasoning effort: chain of thought. Source ID: longbench. Evidence capture ID: curated-longbench-2026-09-02-bfa6246ba625. Source payload hash: bfa6246ba62584289e0204aefc1865fb285a328eef654dcce4382a41b68a3804. Parser version: curated-v1. Imported at: 2026-09-29
- Comparison methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.
LongBench v2: Hard (Show measurement, test conditions, and source)
- Source value
- 49.7
- Score
- 49.7
- Metric
- accuracy with chain of thought
- Unit
- %
- Benchmark version
- v2
- Category
- reasoning
- Task
- hard
- Task label
- Hard
- Direction
- Higher is better
- Comparison cohort
- longbench-v2-cot-hard
- Source type
- official-leaderboard
- Evaluator
- LongBench
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: LongBench. Dataset type: public. Leaderboard date: 2024-11-20. Attempts: pass@1. Reasoning effort: chain of thought. Source ID: longbench. Evidence capture ID: curated-longbench-2026-09-02-185c2fd32a7b. Source payload hash: 185c2fd32a7bedc4dcf0643d2e27095af0b394e9b8a14d95e0c6b0feb36245ac. Parser version: curated-v1. Imported at: 2026-09-29
- Comparison methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.
LongBench v2: Long, 128,000 to 2 million words (Show measurement, test conditions, and source)
- Source value
- 43.5
- Score
- 43.5
- Metric
- accuracy with chain of thought
- Unit
- %
- Benchmark version
- v2
- Category
- reasoning
- Task
- long
- Task label
- Long, 128,000 to 2 million words
- Direction
- Higher is better
- Comparison cohort
- longbench-v2-cot-long
- Source type
- official-leaderboard
- Evaluator
- LongBench
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: LongBench. Dataset type: public. Leaderboard date: 2024-11-20. Attempts: pass@1. Reasoning effort: chain of thought. Source ID: longbench. Evidence capture ID: curated-longbench-2026-09-02-221d7b98957d. Source payload hash: 221d7b98957db163a9e154e9a456439363305e450a9acf020a33ad2fce54bf28. Parser version: curated-v1. Imported at: 2026-09-29
- Comparison methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.
LongBench v2: Medium, 32,000 to 128,000 words (Show measurement, test conditions, and source)
- Source value
- 48.6
- Score
- 48.6
- Metric
- accuracy with chain of thought
- Unit
- %
- Benchmark version
- v2
- Category
- reasoning
- Task
- medium
- Task label
- Medium, 32,000 to 128,000 words
- Direction
- Higher is better
- Comparison cohort
- longbench-v2-cot-medium
- Source type
- official-leaderboard
- Evaluator
- LongBench
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: LongBench. Dataset type: public. Leaderboard date: 2024-11-20. Attempts: pass@1. Reasoning effort: chain of thought. Source ID: longbench. Evidence capture ID: curated-longbench-2026-09-02-d97a751d30a9. Source payload hash: d97a751d30a9e32384a7ce4bdbc341c9dd8f97266490b537b19b7b3026c0fb43. Parser version: curated-v1. Imported at: 2026-09-29
- Comparison methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.
LongBench v2: Overall (Show measurement, test conditions, and source)
- Source value
- 51.4
- Score
- 51.4
- Metric
- accuracy with chain of thought
- Unit
- %
- Benchmark version
- v2
- Category
- reasoning
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- longbench-v2-cot-overall
- Source type
- official-leaderboard
- Evaluator
- LongBench
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: LongBench. Dataset type: public. Leaderboard date: 2024-11-20. Attempts: pass@1. Reasoning effort: chain of thought. Source ID: longbench. Evidence capture ID: curated-longbench-2026-09-02-69320d71d1be. Source payload hash: 69320d71d1be0f7687a4c525303c3abb0ca327d9b7fd521d884acd561a44656d. Parser version: curated-v1. Imported at: 2026-09-29
- Comparison methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.
LongBench v2: Short, up to 32,000 words (Show measurement, test conditions, and source)
- Source value
- 59.6
- Score
- 59.6
- Metric
- accuracy with chain of thought
- Unit
- %
- Benchmark version
- v2
- Category
- reasoning
- Task
- short
- Task label
- Short, up to 32,000 words
- Direction
- Higher is better
- Comparison cohort
- longbench-v2-cot-short
- Source type
- official-leaderboard
- Evaluator
- LongBench
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: LongBench. Dataset type: public. Leaderboard date: 2024-11-20. Attempts: pass@1. Reasoning effort: chain of thought. Source ID: longbench. Evidence capture ID: curated-longbench-2026-09-02-7555ce91dc0b. Source payload hash: 7555ce91dc0bbf8b98caf76a0a7c5981a69b3ee6a8ecc938a1915be2133d61cd. Parser version: curated-v1. Imported at: 2026-09-29
- Comparison methodology
- 503 multiple-choice tasks with the prescribed reasoning prompt.
TutorBench: Overall (Show measurement, test conditions, and source)
- Source value
- 36.12
- Score
- 36.12
- Metric
- weighted rubric score
- Unit
- %
- Benchmark version
- 1,490-task release
- Category
- education
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- scale-tutorbench-overall
- Confidence interval
- 0.96
- Source type
- official-leaderboard
- Evaluator
- Scale AI
- Status
- active
- Retrieved at
- 2026-08-26
- Methodology
- 1,490 tasks with weighted rubrics and Claude 4 Sonnet as judge.. Source model ID: curated/gpt-4o. Source model name: GPT-4o. Provider reported by source: OpenAI. Evaluator: Scale AI. Dataset type: public. Attempts: pass@1. Score display: 36.12% ± 0.96. Source ID: scale-ai. Evidence capture ID: curated-scale-ai-2026-08-26-2d47f3f45c11. Source payload hash: 2d47f3f45c11d5dfb09f242e535afc28a9fd65f5672d1a6a41964d0c2ec8c719. Parser version: curated-v1. Imported at: 2026-09-29
- Comparison methodology
- 1,490 tasks with weighted rubrics and Claude 4 Sonnet as judge.
GPQA Diamond: Few-Shot CoT (Show measurement, test conditions, and source)
- Source value
- 54.04
- Score
- 54.04
- Metric
- accuracy
- Unit
- %
- Benchmark version
- 1
- Category
- reasoning
- Task
- diamond_few_shot_cot
- Task label
- Few-Shot CoT
- Direction
- Higher is better
- Comparison cohort
- vals-gpqa:1:diamond_few_shot_cot
- Source type
- independent-evaluator
- Evaluator
- Vals AI
- Status
- active
- Retrieved at
- 2026-08-11
- Methodology
- Source model ID: openai/gpt-4o-2024-11-20. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384. Latency in seconds: 15.717. Standard error: 3.542. Cost per task in USD: 0.011827. Source ID: vals-ai. Evidence capture ID: vals-gpqa-2026-08-11-d3325290a404. Source payload hash: d3325290a40411054e05ecb86812143465b4d8eee24a0f71f6cf5dfee5b232a1. Parser version: vals-astro-v2. Imported at: 2026-08-11
- Comparison methodology
- Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384
GPQA Diamond: Zero-Shot CoT (Show measurement, test conditions, and source)
- Source value
- 53.535
- Score
- 53.535
- Metric
- accuracy
- Unit
- %
- Benchmark version
- 1
- Category
- reasoning
- Task
- diamond_zero_shot_cot
- Task label
- Zero-Shot CoT
- Direction
- Higher is better
- Comparison cohort
- vals-gpqa:1:diamond_zero_shot_cot
- Source type
- independent-evaluator
- Evaluator
- Vals AI
- Status
- active
- Retrieved at
- 2026-08-11
- Methodology
- Source model ID: openai/gpt-4o-2024-11-20. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384. Latency in seconds: 14.131. Standard error: 3.55. Cost per task in USD: 0.007103. Source ID: vals-ai. Evidence capture ID: vals-gpqa-2026-08-11-d3325290a404. Source payload hash: d3325290a40411054e05ecb86812143465b4d8eee24a0f71f6cf5dfee5b232a1. Parser version: vals-astro-v2. Imported at: 2026-08-11
- Comparison methodology
- Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384
LiveCodeBench v6: Easy (Show measurement, test conditions, and source)
- Source value
- 86.025
- Score
- 86.025
- Metric
- accuracy
- Unit
- %
- Benchmark version
- 1
- Category
- coding
- Task
- easy
- Task label
- Easy
- Direction
- Higher is better
- Comparison cohort
- vals-lcb:1:easy
- Source type
- independent-evaluator
- Evaluator
- Vals AI
- Status
- active
- Retrieved at
- 2026-08-11
- Methodology
- Source model ID: openai/gpt-4o-2024-11-20. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384. Latency in seconds: 2.741. Standard error: 1.932. Cost per task in USD: 0.00269. Source ID: vals-ai. Evidence capture ID: vals-lcb-2026-08-11-0342bcc364e4. Source payload hash: 0342bcc364e4426ffb24589c2752c9225be61fbd26db8deba5bdab9775ce5a75. Parser version: vals-astro-v2. Imported at: 2026-08-11
- Comparison methodology
- Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384
LiveCodeBench v6: Hard (Show measurement, test conditions, and source)
- Source value
- 5.143
- Score
- 5.143
- Metric
- accuracy
- Unit
- %
- Benchmark version
- 1
- Category
- coding
- Task
- hard
- Task label
- Hard
- Direction
- Higher is better
- Comparison cohort
- vals-lcb:1:hard
- Source type
- independent-evaluator
- Evaluator
- Vals AI
- Status
- active
- Retrieved at
- 2026-08-11
- Methodology
- Source model ID: openai/gpt-4o-2024-11-20. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384. Latency in seconds: 5.656. Standard error: 1.181. Cost per task in USD: 0.005244. Source ID: vals-ai. Evidence capture ID: vals-lcb-2026-08-11-0342bcc364e4. Source payload hash: 0342bcc364e4426ffb24589c2752c9225be61fbd26db8deba5bdab9775ce5a75. Parser version: vals-astro-v2. Imported at: 2026-08-11
- Comparison methodology
- Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384
LiveCodeBench v6: Medium (Show measurement, test conditions, and source)
- Source value
- 39.164
- Score
- 39.164
- Metric
- accuracy
- Unit
- %
- Benchmark version
- 1
- Category
- coding
- Task
- medium
- Task label
- Medium
- Direction
- Higher is better
- Comparison cohort
- vals-lcb:1:medium
- Source type
- independent-evaluator
- Evaluator
- Vals AI
- Status
- active
- Retrieved at
- 2026-08-11
- Methodology
- Source model ID: openai/gpt-4o-2024-11-20. Source model name: GPT-4o. Provider reported by source: OpenAI. Model snapshot: gpt-4o-2024-11-20. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-08-09. Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384. Latency in seconds: 4.643. Standard error: 2.494. Cost per task in USD: 0.004063. Source ID: vals-ai. Evidence capture ID: vals-lcb-2026-08-11-0342bcc364e4. Source payload hash: 0342bcc364e4426ffb24589c2752c9225be61fbd26db8deba5bdab9775ce5a75. Parser version: vals-astro-v2. Imported at: 2026-08-11
- Comparison methodology
- Reasoning effort: high. Temperature: 1. Maximum output tokens: 16384
Page 1 of 14
Our tests
SVG short test
One SVG prompt and fixed output settings. Captures are shown only with 0 reported reasoning tokens or documented lack of reasoning controls. Identical test conditions do not make the models' internal compute identical.
Earlier tests
Pelican on a bicycle
A community classic for free-form SVG drawing.
GPT-4o
Reasoning (requested): not specified
Current attempt
- Run date
- Oct 1, 2026
- API provider
- openrouter
- Requested API model
- openai/gpt-4o
- API model reported in response
- openai/gpt-4o
- Reasoning setting
- not specified
- Pinned endpoint
- OpenAI | openai/gpt-4o
- Temperature
- Provider default
- Token limit including reasoning
- 8,192 tokens
- Test protocol
- community-visual-tests-v1
- Technical checks
- Passed, not a quality rating
- Response time
- 7.8 s
Visually reviewed. This individual test does not establish overall model quality.
Prompt and test conditions
Generate an SVG of a pelican riding a bicycle
Token limit including reasoning: 8,192 tokens
Task origin (Simon Willison)Popularity
39.4 / 100
Predecessor and successor
Sources and data date
| Type | Evidence and data date |
|---|---|
| Speed | OpenAI, Nov. Output speed after the first streaming chunk, not total response time. Classified as deprecated by Artificial Analysis. Artificial Analysis maintains only the default 10K-input workload there. Retrieved September 8, 2026. |
| Price | Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs. |
| Input | Documented modalities. Retrieved September 8, 2026. |
| Output | Documented modalities. Retrieved September 8, 2026. |
| Research status | Research date September 8, 2026. 6 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not found in the inspected sources | independent-benchmarks Original wording The recorded GPT-4o benchmark rows are official vendor evaluations; no independent evaluator row was merged. |
| Not found in the inspected sources | parameters-billions Original wording OpenAI does not publish parameter counts. |
| Additional source | Model metadata |
| Additional source | Context window data |
| Additional source | LongBench (retrieved September 2, 2026) |
| Additional source | Scale AI (retrieved August 26, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | Vals AI (retrieved August 11, 2026; September 29, 2026) |
| Additional source | OpenAI (retrieved September 8, 2026) · Hello GPT-4o · Editorial description reviewed October 4, 2026 |
| Additional source | OpenAI (retrieved September 8, 2026) |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |
| Additional source | Introducing GPT-4.1 in the API (2026-09-19) |