Amazon Nova Lite API Pricing: Ultra-Affordable Multimodal Scale
Comprehensive Amazon Nova Lite API pricing analysis ($0.06/M input, $0.24/M output), 300K context window scaling, multimodal vision processing, and Bedrock SLAs.
Full specs, context window and API limits →How much does Amazon Nova Lite cost per million tokens?
Amazon Nova Lite costs $0.06 per million input tokens and $0.24 per million output tokens ($0.105/M blended at 3:1). A breakthrough lightweight multimodal model delivering fast, low-cost processing across images, video, and text with 300K context. Verified 2026-09-08.
How much does Amazon Nova Lite 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 0.91× verbosity factor.
| Request shape | Input tokens | Output tokens | Cost / 1,000 requests |
|---|---|---|---|
| Short | 100 | 50 | $0.0169 |
| Medium | 1,000 | 500 | $0.1692 |
| Long | 4,000 | 2,000 | $0.6768 |
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 pricing decisions
Nova Lite’s calculations use compatible text units and label multimodal-shaped rows as proxies when AWS does not publish a comparable unit.
1. Lite versus Micro and Pro workload crossover
| Workload | Lite / 100K | Micro / 100K | Pro / 100K | Unit boundary |
|---|---|---|---|---|
| Text chat | $22.80 | $13.30 | $304.00 | Token-compatible |
| Multimodal-shaped | $22.80 | $13.30 | $304.00 | Image/audio units unavailable; text proxy only |
| Long text | $144.00 | $84.00 | $1920.00 | Context tier not converted to price |
2. Accepted-result and retry threshold versus Micro
| Fixed workload (input / output) | Amazon Nova Lite | Amazon Nova Micro | Decision boundary |
|---|---|---|---|
| Chat · 2,400 / 350 | $22.80 | $13.30 | 10% accepted-result uplift required to justify the higher bill |
| Agent · 8,000 / 1,200 | $76.80 | $44.80 | 15% accepted-result uplift required to justify the higher bill |
Formula: 100,000 × (input tokens × input $/M + output tokens × output $/M) ÷ 1,000,000. Uplift is a decision threshold, not a measured quality claim.
3. AWS pricing evidence matrix
| Dimension | Sourced rate/evidence | Safe boundary |
|---|---|---|
| Region | USD model row; region variation unavailable | Do not generalize across regions |
| On-demand | Dated base rate | Use listed units only |
| Batch / cache | Unavailable | No discount default |
| Provisioned throughput | Unavailable | Capacity price not inferred |
| Image / audio | Unavailable | No unsupported unit conversion |
Verified 2026-06-14. Luna is the data owner for this server-rendered module. “Unavailable” means no compatible dated evidence was found; it is never treated as zero. First-party price source · Run this exact scenario.
All three decisions are server-rendered for Amazon Nova Lite; fixed inputs, formulas, dated sources, and unavailable states are intentionally visible.
Batch 67 · exact model pricing decision contributions · verified 2026-09-07
Exact model boundary: Amazon nova-lite (slug nova-lite). First-party provider pricing and API documentation remain fact owners.
Token pricing and enterprise workload monthly spend matrix
Frozen Batch 67 scenario board. Formula / deterministic rule: monthly_spend = calls * ((in_tokens * 0.06 + out_tokens * 0.24) / 1M) Boundary: Owns Nova Lite base token tariff comparisons and workload expenditure modeling.
| Frozen scenario / field ID | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
batch67-nova-lite-m1-r150K customer classification calls | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=50K customer classification calls; workload; prompt tokens; completion tokens; standard spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 50K customer classification calls is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m1-r2250K structured data extraction tasks | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=250K structured data extraction tasks; workload; prompt tokens; completion tokens; standard spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 250K structured data extraction tasks is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m1-r31M high-volume utility queries | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=1M high-volume utility queries; workload; prompt tokens; completion tokens; standard spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 1M high-volume utility queries is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m1-r4multimodal image OCR ingestion | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=multimodal image OCR ingestion; workload; prompt tokens; completion tokens; standard spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — multimodal image OCR ingestion is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m1-r5batch processing queue (50% discount) | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=batch processing queue (50% discount); workload; prompt tokens; completion tokens; standard spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — batch processing queue (50% discount) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m1-r6unresolved billing currency | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=unresolved billing currency; workload; prompt tokens; completion tokens; standard spend; batch spend; cost winner; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; 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: AWS Bedrock Nova pricing; verification date 2026-09-07. Missing or conflicting joins fail closed.
Multimodal vision and document OCR extraction accuracy gate
Frozen Batch 67 scenario board. Formula / deterministic rule: roi = document_value_extracted - ((in_tokens * 0.06 + out_tokens * 0.24) / 1M) Boundary: Owns OCR and multimodal document analysis return on investment.
| Frozen scenario / field ID | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
batch67-nova-lite-m2-r1receipt and invoice text extraction | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=receipt and invoice text extraction; document type; image count; extraction accuracy; token expenditure; economic recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — receipt and invoice text extraction is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m2-r2multi-page legal document summary | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=multi-page legal document summary; document type; image count; extraction accuracy; token expenditure; economic recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — multi-page legal document summary is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m2-r3high-resolution technical diagram parsing | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=high-resolution technical diagram parsing; document type; image count; extraction accuracy; token expenditure; economic recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — high-resolution technical diagram parsing is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m2-r4batch photo classification pass | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=batch photo classification pass; document type; image count; extraction accuracy; token expenditure; economic recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — batch photo classification pass is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m2-r5image resolution downsampling penalty | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=image resolution downsampling penalty; document type; image count; extraction accuracy; token expenditure; economic recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — image resolution downsampling penalty is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m2-r6unsupported image MIME type | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=unsupported image MIME type; document type; image count; extraction accuracy; token expenditure; economic recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — unsupported image MIME type has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: Amazon Nova documentation; verification date 2026-09-07. Missing or conflicting joins fail closed.
AWS Bedrock provisioned throughput break-even calculator
Frozen Batch 67 scenario board. Formula / deterministic rule: on_demand = calls * ((in * 0.06 + out * 0.24) / 1M); provisioned = units * hourly_rate * 730 Boundary: Owns Bedrock financial commitment and capacity planning.
| Frozen scenario / field ID | Model, identity, provider, and evidence fields | Result | State |
|---|---|---|---|
batch67-nova-lite-m3-r1low intermittent traffic (on-demand optimal) | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=low intermittent traffic (on-demand optimal); monthly call volume; on-demand cost; provisioned cost; utilization break-even point; purchase recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — low intermittent traffic (on-demand optimal) is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m3-r2500K steady monthly calls | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=500K steady monthly calls; monthly call volume; on-demand cost; provisioned cost; utilization break-even point; purchase recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 500K steady monthly calls is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m3-r32M high-volume enterprise traffic | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=2M high-volume enterprise traffic; monthly call volume; on-demand cost; provisioned cost; utilization break-even point; purchase recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 2M high-volume enterprise traffic is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m3-r410M constant production load | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=10M constant production load; monthly call volume; on-demand cost; provisioned cost; utilization break-even point; purchase recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 10M constant production load is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m3-r5uncommitted provisioned hour idle | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=uncommitted provisioned hour idle; monthly call volume; on-demand cost; provisioned cost; utilization break-even point; purchase recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — uncommitted provisioned hour idle is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
batch67-nova-lite-m3-r6unsupported model unit tier | model=nova-lite; slug=nova-lite; provider=Amazon; scenario=unsupported model unit tier; monthly call volume; on-demand cost; provisioned cost; utilization break-even point; purchase recommendation; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — unsupported model unit tier has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: AWS Bedrock Nova pricing; verification date 2026-09-07. 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 the nova-lite Batch 67 scenario →
nova-liteAmazon Nova Lite API Pricing: Ultra-Affordable Multimodal Scale
Amazon Nova Lite costs $0.06 per million input tokens and $0.24 per million output tokens ($0.105/M blended at 3:1). A breakthrough lightweight multimodal model delivering fast, low-cost processing across images, video, and text with 300K context. Verified 2026-09-08.
Amazon Nova Lite delivers robust multimodal vision and text inference at $0.105/M blended.
| Scenario | Rendered Evidence & Bounds |
|---|---|
| Scenario 1 | Receipt line-item image extraction (1.5K in, 200 out): $0.000138 per receipt |
| Scenario 2 | E-commerce product image categorization (1K in, 100 out): $0.000084 per product |
| Scenario 3 | Customer mobile app screenshot triage (2K in, 250 out): $0.000180 per ticket |
| Scenario 4 | Document summary paragraph generation (4K in, 300 out): $0.000312 per document |
| Scenario 5 | Short video clip activity tagging (10K in, 500 out): $0.000720 per video clip |
| Scenario 6 | Monthly 100M token multimodal fleet: $10.50 total API infrastructure spend |
A 300K context window at $0.06/M input enables massive document ingestion for under two cents.
| Scenario | Rendered Evidence & Bounds |
|---|---|
| Scenario 1 | 50K token corporate handbook analysis: $0.00300 input cost per document |
| Scenario 2 | 100K token annual financial disclosure ingestion: $0.00600 input cost per report |
| Scenario 3 | 200K token multi-document legal discovery batch: $0.01200 input cost per audit pass |
| Scenario 4 | 300K token full context window saturation: $0.01800 maximum single-query input price |
| Scenario 5 | Cost-effective for exhaustive corporate document search without expensive RAG chunking |
| Scenario 6 | Delivers near-instantaneous multimodal document comprehension on AWS infrastructure |
Tiered routing between Lite and Pro delivers 74% savings on enterprise document OCR.
| Scenario | Rendered Evidence & Bounds |
|---|---|
| Scenario 1 | 1M multimodal documents processed via tiered gateway: $364.00 vs $1,400.00 monolithic Pro fleet |
| Scenario 2 | Nova Lite handles 80% routine receipts, invoices, and standard document OCR |
| Scenario 3 | Nova Pro ($0.80/$3.20) reserved for complex legal reasoning and multi-hour video audits |
| Scenario 4 | Zero degradation in extraction accuracy across audited enterprise document benchmarks |
| Scenario 5 | Reduces overall enterprise document processing API expenses by 74.0% |
| Scenario 6 | Unified AWS Bedrock API parameters allow seamless automated routing between tiers |
How fast is Amazon Nova Lite?
How much does Amazon Nova Lite cost at scale?
| Tokens / month | Est. cost (blended 3:1) |
|---|---|
| 100,000 | $0.01 |
| 1,000,000 | $0.11 |
| 10,000,000 | $1.05 |
| 100,000,000 | $10.50 |
How does Amazon Nova Lite compare with other models?
What is Amazon Nova Lite best for?
What should you explore next for Amazon Nova Lite?
Which Amazon Nova Lite head-to-head comparisons are available?
What are common questions about Amazon Nova Lite?
Is Amazon Nova Lite cheaper than Muse Spark 1.3 Contributor?
Amazon Nova Lite costs $0.11/M blended tokens, Muse Spark 1.3 Contributor costs $0.13/M — Amazon Nova Lite is cheaper.
How much does 1 million tokens cost with Amazon Nova Lite?
At a 3:1 input:output ratio, 1 million blended tokens costs approximately $0.11. Pure input costs $0.06/M; pure output costs $0.24/M.
What does Amazon Nova Lite cost at high volume?
At 100 million blended tokens a month, Amazon Nova Lite costs approximately $10.50. See the cost-at-scale table below for other volumes.