OpenAI · model passport

OpenAI: GPT-5

gpt-5

Cloud only
Proprietary licence
400,000 context
Released 2025-08-07(curated)

Overall · LMArena: 1,435#94 of 396
Best percentile: 99.0LiveBench
Cohort standing: 65.1median 51.6

What the data says

Benchmarked on 22 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

99 pct — LiveBench (79.3%)
97 pct — MMLU-Pro (Artificial Analysis) (87.1%)
95 pct — LiveCodeBench (Artificial Analysis) (84.6%)

Weakest axes

13.3 pct — HHEM Hallucination (Vectara) (85.3%)

Uneven: 85.7 percentile points between its best and worst axis — 99 pct on LiveBench versus 13.3 pct on HHEM Hallucination (Vectara).

LiveBench99 pct79.3%Verify →
MMLU-Pro (Artificial Analysis)97 pct87.1%Verify →
LiveCodeBench (Artificial Analysis)95 pct84.6%Verify →
SWE-bench Verified (live board)90.8 pct75.6%Verify →
IFBench (Artificial Analysis)88 pct73.1%Verify →
FrontierMath78.3 pct32.4%Verify →
Overall76.5 pct1,435Verify →
Hard prompts74.2 pct1,448Verify →
AIME (math)74 pct91.4%Verify →
Coding71.4 pct1,471Verify →
GPQA Diamond71 pct86.2%Verify →
Instruction following70.5 pct1,410Verify →
Humanity's Last Exam69.6 pct25.3%Verify →
SimpleQA Verified68.3 pct50.1%Verify →
Terminal-Bench63.6 pct49.6%Verify →
SimpleBench63.2 pct56.7%Verify →
Vision Arena50.7 pct1,210Verify →
WebDev Arena44.6 pct1,418Verify →
ARC-AGI-243.8 pct9.9%Verify →
SWE-bench Verified43.8 pct73.6%Verify →
Factuality Arena30.8 pct1,421Verify →
HHEM Hallucination (Vectara)13.3 pct85.3%Verify →
Search ArenaNot benchmarked
Document ArenaNot benchmarked
Agent ArenaNot benchmarked
BFCL v4 (function calling)Not benchmarked
Terminal-Bench 4.0Not benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Use it hosted

API model IDopenai/gpt-5

Context length400,000 tokens

Max output128,000 tokens

Acceptstext, image, file

Input price$1.25 per 1M tokens

Cached input price$0.13 per 1M tokens

Output price$10 per 1M tokens

API parameters this endpoint supports

Include reasoningMax completion tokensMax tokensReasoningReasoning effortResponse formatSeedStructured outputsTool choiceTool callingVerbosity

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"openai/gpt-5","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

gpt-5-high stands at 65.1 (raw mean 67.2) — measured on 22 of 27 axes.
standingtemplateVerify →
gpt-5-high stands at 65.1 versus a cohort median of 51.6 across 146 models released within 180 days — above the peer median.
cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →