OpenAI · via openai · published window from November 3, 2026

gpt-5.4: stability, speed and price

site endpoint id gpt-5.4

Historical measurements. No successful scored check in the past 24 hours. Scores, latency and cost below describe the 24-hour window ending at the last successful check, not current performance. Current rankings exclude this endpoint.

Last successful scored check: November 3, 2026 9:05 PM UTC

Core pass rate—

83.3%

historical: latest measured day

Speed—

1.1s

historical median per canary call · p95 15.0s

Cost—

$0.173

historical observed mean per scored run · probe mix differs by endpoint

Behaviour fingerprint434 runs

Change detected

12,145 responses · last adj. p 0.00067 · Nov 7 2:23 AM UTC

Summary

gpt-5.4 has no successful scored check in the past 24 hours. The last measured day ended at 2026-11-03T21:05:39.028Z; 83.3% of core checks passed. These historical measurements are excluded from current rankings.

During that historical measurement window, median canary latency was 1.1s (p95 15.0s) and the observed mean cost per scored run was $0.173. This does not describe the endpoint’s present performance.

Raven Alpha has 12,145 fingerprint responses over 434 runs in this endpoint’s selected series (1 other partition of history exist and are not shown here); the last recorded state is “change detected” at 2026-11-07T02:23:54.893Z with an adjusted p of 0.00067.

3 incidents on record, 1 open.

Written from the numbers on this page; every figure is in results.json.

Measurement record

Provider: OpenAI
Route: openai
Model requested: gpt-5.4
Model id returned: gpt-5.4-2026-03-05
Scored calls in window: 30
Fingerprint responses: 12,145
Core score change: —
Status: Historical · no recent check
Suite: hard-v2 + starter-v2 · methodology v0.2

Capabilities over time

Collecting dataTrend charts appear once there are two 6-hour periods of canary runs.

Historical measurement window by capability

Structured JSON100.0%

Schema-compliance regressions

Instruction following100.0%

Exact-format instruction drift

Arithmetic, answer only0.0%

Wrong arithmetic, or a model that starts showing its working when told to give only the number

Tool calling100.0%

Tool-use format regressions

Over-refusal100.0%

New safety layers refusing benign requests

Max-tokens respect100.0%

Truncation and limit changes

Incidents

● Opengpt-5.4Confidence: Low

Raven Alpha (behaviour fingerprint): Responses to fixed prompts no longer match this model’s own history (672 recent vs 5082 baseline responses). Biggest shifts: token probabilities on “Name a famous scientist. Reply with only the name.”; token probabilities on “Say hello in a creative way, in one short sentence.”; token probabilities on “all prompts together”

Change began ≈ Oct 29, 1:57 PM UTC · adjusted p 6.7e-4 · still open · Permalink

✓ Resolvedgpt-5.4Confidence: Low

Raven Alpha (behaviour fingerprint): Responses to fixed prompts no longer match this model’s own history (672 recent vs 5950 baseline responses). Biggest shifts: token probabilities on “Name one animal. Reply with one word.”; first token on “Write a haiku about autumn.”; token probabilities on “Give me a one-line motivational quote.”

Change began ≈ Oct 21, 7:10 PM UTC · adjusted p 5.3e-3 · resolved Nov 4 · Permalink

✓ Resolvedgpt-5.4Confidence: Low

Raven Alpha (behaviour fingerprint): Responses to fixed prompts no longer match this model’s own history (672 recent vs 448 baseline responses). Biggest shifts: first token on “List three uses for a brick.”; token probabilities on “Pick a number between 1 and 1000. Reply with only the number.”; token probabilities on “Suggest a name for a pet cat. Reply with only the name.”

Change began ≈ Oct 5, 8:22 PM UTC · adjusted p 2.4e-3 · resolved Oct 24 · Permalink

Raven Alpha: what it answers

40 runs, 1,120 responses, Nov 4 2:35 PM UTC to Nov 7 2:23 AM UTC. Each row is one fixed prompt with no right answer; the detector tracks the distribution of answers, not any single one. Filled cells: that run's most common answer was the overall favourite. Hollow cells: it was something else (normal for prompts with many good answers).

Prompt	Most common answers (share of all responses)	Median tokens
Pick a random integer between 1 and 50. Reply with only the number. int50 · short answer	88%37 10%17 3%27	4
Pick a number between 1 and 1000. Reply with only the number. int1000 · short answer	68%731 8%613 8%647 8 distinct answers in 40 responses	4
Name one color. Reply with one word. color · short answer	100%Blue	4
Name one animal. Reply with one word. animal · short answer	48%Cat 35%Elephant 18%Tiger	4
Name a city in Europe. Reply with only the city name. city · short answer	100%Prague	5
Give me one random English noun. Reply with one word. noun · short answer	90%lantern 10%harbor	5
Flip a coin. Reply with only heads or tails. coin · short answer	78%tails 23%heads	4
Pick a random letter of the alphabet. Reply with only the letter. letter · short answer	100%Q	4
Name one fruit. Reply with one word. fruit · short answer	100%Apple	4
Suggest a name for a pet cat. Reply with only the name. cat · short answer	95%Mochi 5%Luna	6
Name one programming language. Reply with only its name. lang · short answer	100%Python	4
Name a chemical element. Reply with one word. element · short answer	98%Oxygen 3%Helium	5
Give one adjective that describes the ocean. Reply with one word. ocean · short answer	100%vast	4
Name a famous scientist. Reply with only the name. scientist · short answer	80%Albert Einstein 20%Marie Curie	5
Pick a day of the week. Reply with one word. weekday · short answer	95%Thursday 5%Tuesday	4
Name one musical instrument. Reply with one word. instrument · short answer	80%Piano 20%Violin	5
Write one sentence about rain. rain · free text	45%Rain taps softly on the… 23%Rain taps softly on rooftops… 8%Rain falls from clouds and… 11 distinct answers in 40 responses	18
Explain what a variable is in programming, in two sentences. variable · free text	93%A variable is a named… 8%A variable in programming is…	47
List three uses for a brick. brick · free text	45%- Build a wall -… 28%1. Building walls or structures… 18%- Build a wall or… 6 distinct answers in 40 responses	25
*What is 17 23?** mult · free text	100%391…	4
Write one sentence about the sea. sea · free text	38%The sea shimmered under the… 18%The sea stretched to the… 13%The sea stretches to the… 9 distinct answers in 40 responses	17
Complete this sentence: The old house at the end of the street house · free text	100%The old house at the…	28
Describe a cat in exactly five words. catfive · free text	18%Soft, whiskered, curious, graceful, independent… 8%Graceful, curious, whiskered, independent, purring… 8%Soft, curious, whiskered, graceful, independent… 27 distinct answers in 40 responses	14
Give me a one-line motivational quote. quote · free text	40%“Small steps taken consistently can… 18%“Small steps, taken consistently, can… 15%“Small steps every day turn… 12 distinct answers in 40 responses	19
Say hello in a creative way, in one short sentence. hello · free text	63%Greetings, stardust traveler… 5%Hello—consider this a tiny firework… 5%Hello—consider this a tiny parade… 13 distinct answers in 40 responses	10
Write a haiku about autumn. haiku · free text	18%Crisp leaves drift softly Amber… 8%Crisp leaves drift softly down… 8%Crisp leaves drift downward Amber… 26 distinct answers in 40 responses	22
Write the first sentence of a mystery novel. story · free text	88%By the time the church… 8%By the time the clock… 5%By the time the town…	24
Give one tip for sleeping better, in one sentence. tip · free text	100%Keep a consistent sleep schedule…	22

Raw file: gpt-5.4.json (the same reference martingale check compares your endpoint against).

Each category is scored by deterministic checks only (Structured JSON, Tool calling and others). Read the methodology.