Inference
EN
OpenAI

GPT-5 nano

GPT-5 nano is a reasoning model from OpenAI in the GPT-5 family. 19 benchmarks count toward its index, in 6 categories.

  • Closed weights
  • Reasoning
  • Input: text · image · file
  • Released 2025-08
  • Retires 11 Dec 2026
RankedEnough results for an index.
IndexAll categories, out of 100.41.9±6.2Median 51.0
CoverageIndex weight that has results.75%
SpeedOutput tokens per second.81/sMedian 67/s
First tokenSeconds to the first output token.2.5 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.05batch $0.025Median $0.60
Output / 1MUS dollars per 1M output tokens.$0.40batch $0.20 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.400KMedian 262K
EloLMArena rating and rank.1320±7 #237Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
openai/gpt-5-nano
Max outputMaximum output tokens per request.
128K
Knowledge cutoffLast date of training data.
31 May 2024
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
<$0.01
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
11 Dec 2026

Where to run

APIs that serve this model.
5 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
DatabricksCloud$0.05$0.40400K
DigitalOceanCloud$0.05$0.40400K
OpenRouterRouter$0.05$0.40400K
SAP AI CoreCloud$0.05$0.40400K
Snowflake CortexCloud$0.06$0.44272K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.None counted——
CodingCode writing and repair.2 counted37.2185 / 209
ReasoningLogic problems and puzzles.4 counted42.6151 / 227
MultimodalTasks with images and text.1 counted47.5—
KnowledgeFacts and expert knowledge.3 counted37.4212 / 268
MultilingualTasks in many languages.1 counted44.5—
InstructionTasks with strict prompt rules.None counted——
MathMath problems.8 counted45.8107 / 168

Results

19 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
MATH level 5Math94.9% ±0.511 / 88
EuroEval SwedishMultilingual62.3% ±0.529 / 195
MATH 500Math93.8% ±1.115 / 59
EuroEval GermanMultilingual47.0% ±0.651 / 191
EuroEval PolishMultilingual52.8% ±0.556 / 209
EuroEval SpanishMultilingual49.0% ±0.548 / 176
EuroEval DutchMultilingual53.7% ±0.549 / 175
EuroEval ItalianMultilingual53.8% ±0.549 / 174
EuroEval FrenchMultilingual53.9% ±0.570 / 178
EuroEval PortugueseMultilingual49.0% ±0.673 / 182
OTIS Mock AIME 2024-2025Math81.1% ±4.685 / 200
AIMEMath81.2% ±1.345 / 95
GPQA diamond · Epoch AIKnowledge69.4% ±2.8114 / 217
FrontierMath-2025-02-28-PrivateMath8.3% ±1.640 / 69
LiveCodeBenchCoding70.2% ±0.988 / 142
Chess PuzzlesReasoning22.0% ±4.257 / 90
SWE-bench Verified · swebench.comCoding34.8%30 / 46
ARC-AGI-2 (semi-private)Reasoning2.6%93 / 131
MMMU ProMultimodal70.9% ±1.169 / 92
GPQA Diamond · Vals AIKnowledge63.4% ±2.5105 / 135
MMLU ProKnowledge76.1% ±0.4107 / 135
SWE-bench Verified · SWE-bench/experimentsCoding34.8%53 / 65
FrontierMath-Tier-4-2025-07-01-PrivateMath0.0%45 / 55
Mystery Game PuzzlesReasoning8.0% ±2.766 / 80
ARC-AGI-1 (semi-private)Reasoning16.7%107 / 126
FrontierMath-Tiers-1-3-v2-PrivateMath20.0% ±2.478 / 90
FrontierMath-Tier-4-v2-PrivateMath2.4% ±2.465 / 72
SimpleQA VerifiedKnowledge11.7% ±1.087 / 91
FrontierMath-Tier-4-2025-07-01-PublicMath0.0%3 / 36
MedQAHealthcare93.3% ±0.223 / 95
FrontierMath-2025-02-28-PublicMath20.0% ±13.326 / 64
MGSMMultilingual89.3% ±1.947 / 74
TaxEval v2Finance67.4% ±0.9107 / 143
MedScribeHealthcare72.9% ±1.985 / 109
CaseLaw v2Legal52.6% ±1.042 / 53
ARC-AGI-2 (public eval)Reasoning0.3%99 / 122
MortgageTaxFinance53.6% ±1.080 / 97
ARC-AGI-1 (public eval)Reasoning29.7%98 / 116
MedCodeHealthcare30.4% ±1.995 / 106
LegalBenchLegal50.1% ±0.5147 / 149

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYVals AINo licence statedEuroEvalMITSWE-bench teamNo licence statedARC Prize FoundationTerms restrict reusemodels.devMIT

History

5 releases
Index42.9 → 41.9
23 Oct 202612 Nov 2026
Output / 1M$0.40 → $0.40
23 Oct 202612 Nov 2026

Questions

How much does GPT-5 nano cost?
$0.05 per 1M input tokens, $0.40 per 1M output tokens.
How good is GPT-5 nano?
Its index is 41.9 out of 100. That is place 199 of 283 ranked models.
What is GPT-5 nano best at?
Math: 45.8 out of 100, place 107 of 168.
How fast is GPT-5 nano?
81 output tokens per second, 2.5 s to the first token.
What is the context window of GPT-5 nano?
400K tokens.
Is GPT-5 nano open source?
No. Its weights are not published.
When was GPT-5 nano released?
August 2025.

Badge

GPT-5 nano on Inference360: place 199 of 283, index 41.9
HTML
<a href="https://inference360.ai/models/gpt-5-nano"><img src="https://inference360.ai/badges/gpt-5-nano.svg" alt="GPT-5 nano on Inference360: place 199 of 283, index 41.9" height="20"></a>
Markdown
[![GPT-5 nano on Inference360: place 199 of 283, index 41.9](https://inference360.ai/badges/gpt-5-nano.svg)](https://inference360.ai/models/gpt-5-nano)

Compare with

GLM-4.641.8GPT-OSS 20B41.7Qwen3 235B 250742.3GPT-5 mini48.4GPT-555.0

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from OpenAI

All models
GPT-5 (high)53.6GPT-5 (medium)53.0GPT-5 mini48.4GPT-555.0GPT-OSS 120B43.4