GPT-5.4 nano (xhigh) API Pricing

OpenAI · released 2026-03-17 · merges 3 reasoning-effort tiers

Benchmarks shown for the highest-scoring reasoning tier. Merged tiers: GPT-5.4 nano (xhigh), GPT-5.4 nano (medium), GPT-5.4 nano (Non-Reasoning).

Pricing

$0.20
input / 1M tokens
0.020¢ / 1K
$1.25
output / 1M tokens
0.125¢ / 1K
$0.463
blended / 1M tokens
0.046¢ / 1K

What it actually costs

WorkloadTokensCost
Short chat reply500 in / 300 out0.0475¢
RAG query4,000 in / 500 out0.143¢
Long document summary50,000 in / 1,000 out$0.0113
Code review of a 2k-line file30,000 in / 2,000 out0.85¢
1,000 chat replies/day for 30 days30,000 × (500 in / 300 out)$14.25

Cost = (input tokens ÷ 1,000,000 × $0.20) + (output tokens ÷ 1,000,000 × $1.25).

Benchmarks

20.7
Intelligence
#123 of 533
56.1
Coding
#55 of 197
Math
Tokens/sec
Time to first token

Similar intelligence, lower price

ModelIntelligenceBlended $/1M
Hy3-preview (Reasoning)22.7$0.10
MiMo-V2-Flash (Reasoning)20.8$0.15
Grok 4 Fast (Reasoning)17.9$0.275
DeepSeek V3.2 (Reasoning)21.5$0.315
Step 3.7 Flash19.5$0.438

Leaderboard neighbours

ModelIntelligenceBlended $/1M
MiniMax-M2.120.9$0.525
MiMo-V2-Flash (Reasoning)20.8$0.15
Claude 4.5 Sonnet (Reasoning)20.7$6.00
GPT-5.4 nano (xhigh)20.7$0.463
Claude 4 Opus (Reasoning)20.6$30.00
GPT-5 mini (medium)20.6$0.688
K2 Horizon 7B20.6

More from OpenAI

ModelIntelligenceBlended $/1M
GPT-6 Astra (max)52.7$20.00
GPT-5.6 Sol (max)47.0$8.00
GPT-5.6 Terra (max)42.1$4.50
GPT-5.4 (xhigh)39.0$5.625
GPT-5.5 (xhigh)38.4$11.25
GPT-5.6 Luna (max)37.3$0.45
GPT-5.3 Codex (xhigh)32.5$4.813
GPT-5.2 (xhigh)30.4$4.813

Count tokens & estimate cost for GPT-5.4 nano (xhigh) See GPT-5.4 nano (xhigh) on the leaderboard

Source: Artificial Analysis · last synced 2026-09-22

FAQ

Who develops GPT-5.4 nano (xhigh)?

GPT-5.4 nano (xhigh) is developed by OpenAI, released 2026-03-17.

How much does GPT-5.4 nano (xhigh) cost per million tokens?

GPT-5.4 nano (xhigh) costs $0.20 per 1M input tokens and $1.25 per 1M output tokens ($0.463 blended), per Artificial Analysis's official pricing data.

How does GPT-5.4 nano (xhigh) rank on benchmarks?

GPT-5.4 nano (xhigh) scores 20.7 on the Artificial Analysis Intelligence Index, ranking #123 of 533 models we track.

Is there a cheaper model with similar intelligence to GPT-5.4 nano (xhigh)?

Yes — Hy3-preview (Reasoning) scores similar intelligence (22.7) at $0.10/1M blended, versus $0.463/1M for GPT-5.4 nano (xhigh).

Where does this data come from?

Pricing and benchmark data for GPT-5.4 nano (xhigh) comes from Artificial Analysis (artificialanalysis.ai), synced daily. Last synced 2026-09-22.