OpenAI GPT-5.4 Nano API Pricing: Rapid Edge Triage & Classification
Examine OpenAI GPT-5.4 Nano API pricing ($0.20/M input, $1.25/M output), ultra-low latency intent classification, sub-second routing, and batch savings.
How much does GPT-5.4 Nano cost per million tokens?
OpenAI GPT-5.4 Nano costs $0.20 per million input tokens and $1.25 per million output tokens ($0.4625/M blended at 3:1). Engineered for high-throughput classification, sub-100ms intent routing, and real-time guardrails. Verified 2026-09-08.
How much does GPT-5.4 Nano cost per 1,000 requests?
Computed from generated token pricing. Each row assumes the listed input and output tokens per request; output is adjusted by this model's measured 0.78× verbosity factor.
| Request shape | Input tokens | Output tokens | Cost / 1,000 requests |
|---|---|---|---|
| Short | 100 | 50 | $0.0688 |
| Medium | 1,000 | 500 | $0.6875 |
| Long | 4,000 | 2,000 | $2.7500 |
Formula: ((input price × input tokens) + (output price × output tokens × verbosity factor)) ÷ 1,000,000 × 1,000. Assumptions: short 100/50, medium 1,000/500, long 4,000/2,000 input/output tokens per request. Verbosity run: 2026-06-16T20:31:30.728Z.
Three model-specific legacy pricing decisions
GPT-5.4 Nano owns high-volume classification, routing, and short-generation economics; image, latency, and quota units are not substituted.
1. Fixed model-specific workload bills
| Workload | Input / output | 100K requests | Evidence boundary |
|---|---|---|---|
| Classification | 300 / 30 | $9.75 | 100K / 1M / 10M ladder |
| Routing | 500 / 80 | $20.00 | 100K / 1M / 10M ladder |
| Short generation | 1,000 / 500 | $82.50 | Text tokens |
Formula: requests × (input tokens × input $/M + output tokens × output $/M × output expansion) ÷ 1,000,000. Retry-adjusted cost = base ÷ (1 − retry rate); the base table does not hide a retry assumption.
2. Nano → Luna migration premium
| Fixed shape | GPT-5.4 Nano | GPT-6 Luna | Numeric decision boundary |
|---|---|---|---|
| Classification · 300 / 30 | $10.26 | $4.69 | 25% accepted-result uplift required after fixed retry assumptions (5% → 4%) |
| Routing · 500 / 80 | $21.05 | $9.38 | 25% accepted-result uplift required after fixed retry assumptions (5% → 4%) |
This is a cost-per-accepted-result threshold, not a measured quality claim. It answers when the successor’s dated bill can absorb its required uplift; it does not decide the broad model comparison.
3. Budget-impact and missing-input ledger
| Traffic / evidence | Result | Safe treatment |
|---|---|---|
| 100K requests | $10.26 | Fixed first workload; text-token units |
| 1M requests | $97.50 | Linear token spend only; no volume discount inferred |
| 10M requests | $975.00 | Budget exposure; quota and latency remain unavailable |
| Latency and quota: Unavailable | Unavailable | Do not infer, zero-price, or import a neighboring model’s mechanic |
| Cache and batch: Unavailable | Unavailable | Do not infer, zero-price, or import a neighboring model’s mechanic |
| Shutdown date: Unavailable; no indefinite-availability inference | Unavailable | Do not infer, zero-price, or import a neighboring model’s mechanic |
| Lifecycle | legacy; no sourced announcement date | No sourced shutdown date; revalidate before current claims |
Nano-to-Luna premium by workload shape
| Workload | Nano cost | Luna cost | Absolute premium | Premium | Accepted-result uplift required |
|---|---|---|---|---|---|
| Classification | $9.75 | $4.50 | $-5.2500 | -53.8% | -53.8% |
| Routing | $20.00 | $9.00 | $-11.0000 | -55.0% | -55.0% |
| Short generation | $82.50 | $35.00 | $-47.5000 | -57.6% | -57.6% |
Premium = Luna cost − Nano cost; percentage premium and required accepted-result uplift = (Luna ÷ Nano − 1) × 100 at each fixed shape.
Price verified 2026-04-06; lifecycle verified 2026-08-14. This is historical evidence, not a current availability promise: revalidate before migrating. “Unavailable” means no compatible dated evidence was found; it is never treated as zero. First-party price source · lifecycle source · Test this model in All AI Ask.
Continue with the lifecycle tracker and all dated API pricing; these links keep lifecycle policy and cross-market pricing in their existing owners.
All three decisions below are computed for GPT-5.4 Nano; fixed inputs, formulas, dated provenance, successor boundary, lifecycle state, and missing-data treatment remain visible.
Exact-model pricing guide · verified 2026-09-08
Exact model boundary: OpenAI gpt-5.4-nano (slug gpt-5-4-nano). First-party provider pricing and API documentation remain fact owners.
Ultra-budget token pricing and high-volume classification matrix
Frozen scenario board. Formula / deterministic rule: monthly_spend = calls * ((in_tokens * 0.20 + out_tokens * 1.25) / 1M) Boundary: Owns GPT-5.4 Nano token tariff modeling and high-volume classification spend.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| 100K intent routing classifications | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=100K intent routing classifications; workload; prompt tokens; completion tokens; standard spend; cached spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 100K intent routing classifications is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 500K customer inquiry triage calls | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=500K customer inquiry triage calls; workload; prompt tokens; completion tokens; standard spend; cached spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 500K customer inquiry triage calls is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 2M automated content moderation passes | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=2M automated content moderation passes; workload; prompt tokens; completion tokens; standard spend; cached spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 2M automated content moderation passes is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| prompt caching reuse (50% input discount) | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=prompt caching reuse (50% input discount); workload; prompt tokens; completion tokens; standard spend; cached spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — prompt caching reuse (50% input discount) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| batch processing API queue (50% discount) | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=batch processing API queue (50% discount); workload; prompt tokens; completion tokens; standard spend; cached spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — batch processing API queue (50% discount) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unresolved billing currency | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=unresolved billing currency; workload; prompt tokens; completion tokens; standard spend; cached spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — unresolved billing currency has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: OpenAI API pricing; verification date 2026-09-08. Missing or conflicting joins fail closed.
Sub-second latency SLA and intent classification turnaround audit
Frozen scenario board. Formula / deterministic rule: turnaround = ttft + (tokens_out / tps); optimized for lowest TTFT Boundary: Owns turnaround time benchmarks and user experience responsiveness.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| real-time intent classification (<80ms) | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=real-time intent classification (<80ms); task; response SLA; TTFT; generation speed; SLA compliance; responsiveness champion; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — real-time intent classification (<80ms) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| customer support routing ticket (<150ms) | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=customer support routing ticket (<150ms); task; response SLA; TTFT; generation speed; SLA compliance; responsiveness champion; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — customer support routing ticket (<150ms) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| content moderation guardrail pass (<120ms) | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=content moderation guardrail pass (<120ms); task; response SLA; TTFT; generation speed; SLA compliance; responsiveness champion; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — content moderation guardrail pass (<120ms) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| high-concurrency request surge | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=high-concurrency request surge; task; response SLA; TTFT; generation speed; SLA compliance; responsiveness champion; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — high-concurrency request surge is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| network transit buffer | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=network transit buffer; task; response SLA; TTFT; generation speed; SLA compliance; responsiveness champion; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — network transit buffer is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unmeasured speed fixture | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=unmeasured speed fixture; task; response SLA; TTFT; generation speed; SLA compliance; responsiveness champion; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — unmeasured speed fixture is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
First-party provenance: OpenAI API documentation; verification date 2026-09-08. Missing or conflicting joins fail closed.
Two-tier triage architecture: Nano triage vs Mini execution
Frozen scenario board. Formula / deterministic rule: blended_cost = nano_volume * nano_cost + mini_volume * mini_cost Boundary: Owns architectural routing between fast Nano triage and deeper Mini execution.
| Frozen scenario | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
| 100% GPT-5.4 Nano baseline ($0.20/$1.25) | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=100% GPT-5.4 Nano baseline ($0.20/$1.25); escalation mix; total monthly queries; blended spend; cost savings vs pure Mini; accuracy retention; routing recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 100% GPT-5.4 Nano baseline ($0.20/$1.25) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 90% Nano triage / 10% Mini escalation | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=90% Nano triage / 10% Mini escalation; escalation mix; total monthly queries; blended spend; cost savings vs pure Mini; accuracy retention; routing recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 90% Nano triage / 10% Mini escalation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 80% Nano triage / 20% Mini escalation | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=80% Nano triage / 20% Mini escalation; escalation mix; total monthly queries; blended spend; cost savings vs pure Mini; accuracy retention; routing recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 80% Nano triage / 20% Mini escalation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 50% Nano triage / 50% Mini escalation | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=50% Nano triage / 50% Mini escalation; escalation mix; total monthly queries; blended spend; cost savings vs pure Mini; accuracy retention; routing recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 50% Nano triage / 50% Mini escalation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 100% direct GPT-5.4 Mini execution | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=100% direct GPT-5.4 Mini execution; escalation mix; total monthly queries; blended spend; cost savings vs pure Mini; accuracy retention; routing recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — 100% direct GPT-5.4 Mini execution is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unresolved routing confidence threshold trigger | model=gpt-5.4-nano; slug=gpt-5-4-nano; provider=OpenAI; scenario=unresolved routing confidence threshold trigger; escalation mix; total monthly queries; blended spend; cost savings vs pure Mini; accuracy retention; routing recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-08; measurement versus assumption=explicit | Unavailable — unresolved routing confidence threshold trigger has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: OpenAI API pricing; verification date 2026-09-08. Missing or conflicting joins fail closed.
Method and limitations: formulas are deterministic; observed and assumed inputs are labeled; no missing provider, host, account, region, realm, alias, snapshot, revision, weight, artifact, control, tool, modality, workload, rate-period, timestamp, or result is transferred. Run this scenario →
GPT-5.4 Nano pricing evidence
Source-backed rate shape: input=$0.200000/M; output=$1.250000/M. Historical registry snapshot; revalidate before any current-price claim. Nano dated routing economics only; Luna pricing, cheapest rankings, provider limits, and calculator scenarios retain ownership.
Module 1 of 3
Tagging/routing request ladder
Novel contribution boundary: Owns dated Nano bills for fixed tagging and routing request shapes. Formula / deterministic rule: spend = requests × (inputTokens × inputRate + outputTokens × outputRate) / 1,000
| Scenario | Exact model, provider, and fixed inputs | Result | State |
|---|---|---|---|
| 100K tagging requests | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; requests=100,000; inputTokens/request=300; outputTokens/request=80; perRequest=$0.000160; fixed inputs only; measurement versus assumption=explicit | Deterministic spend = $16.000000 for 100,000 requests; no outcome or benchmark asserted. | VERIFIED_DETERMINISTIC — arithmetic only |
| 1M routing requests | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; requests=1,000,000; inputTokens/request=600; outputTokens/request=120; perRequest=$0.000270; fixed inputs only; measurement versus assumption=explicit | Deterministic spend = $270.000000 for 1,000,000 requests; no outcome or benchmark asserted. | VERIFIED_DETERMINISTIC — arithmetic only |
| 10M compact event routes | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; requests=10,000,000; inputTokens/request=1,000; outputTokens/request=200; perRequest=$0.000450; fixed inputs only; measurement versus assumption=explicit | Deterministic spend = $4500.000000 for 10,000,000 requests; no outcome or benchmark asserted. | VERIFIED_DETERMINISTIC — arithmetic only |
First-party registry provenance: https://openai.com/api/pricing; registry provider openai; verified 2026-04-06. Revalidate before any current-price claim.
Module 2 of 3
Nano→Luna multiplier and retry break-even
Novel contribution boundary: Owns dated price multiplier arithmetic with explicit retry input; accepted-result and quota observations are unavailable. Formula / deterministic rule: multiplier = Luna attempt-plus-retry cost / Nano attempt cost; accepted quality = Unavailable
| Scenario | Exact model, provider, and fixed inputs | Result | State |
|---|---|---|---|
| Short route; 10% retry share; 100 retry-input tokens | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; ownerAttempt=$0.000160; comparisonAttempt=$0.000780; retryInputTokens=100; retryShare=10%; fixed inputs only; measurement versus assumption=explicit | Deterministic comparison cost with explicit retry input = $0.000790; required accepted-result rate = 493.75%; observed acceptance/quality = Unavailable. | VERIFIED_DETERMINISTIC — arithmetic only |
| Medium route; 20% retry share; 250 retry-input tokens | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; ownerAttempt=$0.000270; comparisonAttempt=$0.001320; retryInputTokens=250; retryShare=20%; fixed inputs only; measurement versus assumption=explicit | Deterministic comparison cost with explicit retry input = $0.001370; required accepted-result rate = 507.41%; observed acceptance/quality = Unavailable. | VERIFIED_DETERMINISTIC — arithmetic only |
| Long route; 30% retry share; 500 retry-input tokens | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; ownerAttempt=$0.000450; comparisonAttempt=$0.002200; retryInputTokens=500; retryShare=30%; fixed inputs only; measurement versus assumption=explicit | Deterministic comparison cost with explicit retry input = $0.002350; required accepted-result rate = 522.22%; observed acceptance/quality = Unavailable. | VERIFIED_DETERMINISTIC — arithmetic only |
First-party registry provenance: https://openai.com/api/pricing; registry provider openai; verified 2026-04-06. Revalidate before any current-price claim.
Module 3 of 3
Migration-risk budget separating sourced rates from unavailable inputs
Novel contribution boundary: Owns dated Nano rate arithmetic; quality, latency, cache, batch, quota, and shutdown inputs are not inferred. Formula / deterministic rule: riskBudget = sourced token spend; quality/latency/cache/batch/quota/shutdown = Unavailable
| Scenario | Exact model, provider, and fixed inputs | Result | State |
|---|---|---|---|
| Sourced dated input/output rates | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; FAIL CLOSED — no sourced observation or provider guarantee; measurement versus assumption=explicit | Unavailable — the registry does not publish cache evidence for this exact model. | FAIL CLOSED — no sourced observation or provider guarantee |
| Quota, cache, or batch mechanics | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; FAIL CLOSED — no sourced observation or provider guarantee; measurement versus assumption=explicit | Unavailable — the registry does not publish batch evidence for this exact model. | FAIL CLOSED — no sourced observation or provider guarantee |
| Latency, quality, or shutdown evidence | model=gpt-5.4-nano; provider=openai; registryRates=input=$0.200000/M; output=$1.250000/M; comparisonModel=gpt-5.6-luna; FAIL CLOSED — no sourced observation or provider guarantee; measurement versus assumption=explicit | Unavailable — the registry does not publish context evidence for this exact model. | FAIL CLOSED — no sourced observation or provider guarantee |
First-party registry provenance: https://openai.com/api/pricing; registry provider openai; verified 2026-04-06. Revalidate before any current-price claim.
How fast is GPT-5.4 Nano?
How much does GPT-5.4 Nano cost at scale?
| Tokens / month | Est. cost (blended 3:1) |
|---|---|
| 100,000 | $0.05 |
| 1,000,000 | $0.46 |
| 10,000,000 | $4.63 |
| 100,000,000 | $46.25 |
How does GPT-5.4 Nano compare with other models?
What should you explore next for GPT-5.4 Nano?
What are common questions about GPT-5.4 Nano?
Is GPT-5.4 Nano cheaper than Codestral?
GPT-5.4 Nano costs $0.46/M blended tokens, Codestral costs $0.45/M — Codestral is cheaper.
How much does 1 million tokens cost with GPT-5.4 Nano?
At a 3:1 input:output ratio, 1 million blended tokens costs approximately $0.46. Pure input costs $0.20/M; pure output costs $1.25/M.
What does GPT-5.4 Nano cost at high volume?
At 100 million blended tokens a month, GPT-5.4 Nano costs approximately $46.25. See the cost-at-scale table below for other volumes.