Skip to content
aiqbench

OpenAI · model passport

OpenAI: GPT-5.4

gpt-5-4
  • Cloud only
  • Proprietary licence
  • 1,050,000 context
  • Released 2026-09-13(first seen on LMArena)
20/27axes covered
Overall · LMArena
1,475#30 of 396
Best percentile
97.7Terminal-Bench
Cohort standing
78.2median 60.3

What the data says

Benchmarked on 20 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

  • 97.7 pct — Terminal-Bench (81.8%)
  • 97.3 pct — LiveBench (78.0%)
  • 95.9 pct — Factuality Arena (1,484)

Uneven: 45.3 percentile points between its best and worst axis — 97.7 pct on Terminal-Bench versus 52.4 pct on SimpleQA Verified.

  • Terminal-Bench97.7 pct81.8%Verify →
  • LiveBench97.3 pct78.0%Verify →
  • Factuality Arena95.9 pct1,484Verify →
  • FrontierMath95.7 pct47.6%Verify →
  • Overall92.7 pct1,475Verify →
  • Coding92.6 pct1,520Verify →
  • GPQA Diamond92.3 pct93.3%Verify →
  • Instruction following92.2 pct1,471Verify →
  • Hard prompts91.7 pct1,496Verify →
  • IFBench (Artificial Analysis)89.8 pct73.9%Verify →
  • Vision Arena87.3 pct1,285Verify →
  • AIME (math)85.5 pct97.8%Verify →
  • Humanity's Last Exam82.6 pct36.2%Verify →
  • ARC-AGI-280.9 pct74.0%Verify →
  • SWE-bench Verified78.1 pct76.9%Verify →
  • HHEM Hallucination (Vectara)70.5 pct93.0%Verify →
  • Document Arena65.8 pct1,470Verify →
  • WebDev Arena55.4 pct1,464Verify →
  • Agent Arena52.7 pct0.016Verify →
  • SimpleQA Verified52.4 pct45.1%Verify →
  • Search ArenaNot benchmarked
  • SimpleBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not 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-5.4
Context length1,050,000 tokens
Max output128,000 tokens
Acceptstext, image, file
Input price$2.50 per 1M tokens
Cached input price$0.25 per 1M tokens
Output price$15 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-5.4","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-5.4-high stands at 78.2 (raw mean 82.5) — measured on 20 of 27 axes.

    standingtemplateVerify →

  • gpt-5.4-high stands at 78.2 versus a cohort median of 60.3 across 176 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 →