Atlas Graph Explorer

model:gpt-5@current

ModelVersioncompute/models/gpt-5.yaml·Open in Graph →

overview json graph

Attributes

supportsExtendedThinking

false

modelFamilyId

model-family:gpt-5

versionRange

>=5.0.0 <6.0.0

kind

reasoning

contextWindowTokens

400000

costPerMTokInput

1.25

costPerMTokOutput

supportsThinking

true

thinkingEffortLevels

minimal
low
medium
high

supportsThinkingBudgetTokens

false

modalities

text
image

modalityDirections

kind
text
direction
input
kind
image
direction
input
kind
text
direction
output

releaseDate

2025-08-07

lifecycleStatus

lifecycleStatusChangedAt

2025-08-07

knowledgeCutoffDate

2024-10-01

safetyTier

openai-preparedness-high

modelCardUrl

https://openai.com/index/gpt-5-system-card/

supportsFineTuning

false

regions

global

Outgoing edges (19)

realizes1

layer:1-model·LayerModel

served_by1

provider:openai·ModelProviderProductOpenAI

speaks2

model-transport:openai-responses·ModelTransportProtocolOpenAI Responses
model-transport:openai-chat-completions·ModelTransportProtocolOpenAI Chat Completions

supports10

capability:supports-batch-api·CapabilitySupports batch API
capability:parallel-tool-calls·CapabilityParallel tool calls
capability:supports-prompt-caching·CapabilitySupports prompt caching
capability:streaming·CapabilityStreaming
capability:supports-text-streaming·CapabilitySupports text streaming
capability:supports-thinking-streaming·CapabilitySupports thinking streaming
capability:supports-thinking·CapabilitySupports thinking
capability:supports-tool-call-streaming·CapabilitySupports tool-call streaming
capability:supports-tool-use·CapabilitySupports tool use
capability:vision-input·CapabilityVision input

supports_effort_level4

effort-level:minimal·EffortLevelEffort: minimal
effort-level:low·EffortLevelEffort: low
effort-level:medium·EffortLevelEffort: medium
effort-level:high·EffortLevelEffort: high

version_of1

model-family:gpt-5·ModelFamilyGPT-5

Incoming edges (28)

about_subject15

evaluates_target9

eval-run:gpqa.gpt-5.2025-08·EvalRun
eval-run:human-eval.gpt-5.2025-08·EvalRun
eval-run:bfcl.gpt-5.2025-08·EvalRun
eval-run:gpqa-diamond.gpt-5.2025-08·EvalRun
eval-run:human-eval-plus.gpt-5.2025-08·EvalRun
eval-run:livecodebench.gpt-5.2025-08·EvalRun
eval-run:math.gpt-5.2025-08·EvalRun
eval-run:evalplus.gpt-5.2025-08·EvalRun
eval-run:swe-bench-verified.gpt-5.2025-08·EvalRun

has_version1

model-family:gpt-5·ModelFamilyGPT-5

serves1

provider:openai·ModelProviderProductOpenAI

spoken_by2

model-transport:openai-chat-completions·ModelTransportProtocolOpenAI Chat Completions
model-transport:openai-responses·ModelTransportProtocolOpenAI Responses