TheRouter.aiTheRouter.ai
Back to Models

Amazon Nova 2 Lite

amazonamazon/nova-2-lite

Nova 2 Lite is an advanced multimodal reasoning model with 1M context. Dynamically adjusts reasoning depth. Extended thinking on complex problems.

AWS announcement ↗Nova product page ↗Bedrock model card ↗Nova 2 technical report (PDF) ↗

Amazon Nova 2 Lite is the small, agentic reasoning model in Amazon's Nova 2 family, announced in early December 2025 and served through Amazon Bedrock and SageMaker AI. It accepts text, images, documents, and video, with a 1 million-token context window — large enough to take a full codebase or a long-form video transcript as input in a single call.

From a TheRouter operator's perspective, Nova 2 Lite is what you reach for when you need a million-token context at mini-tier price. Extended thinking is off by default — Amazon ships it with three opt-in budget levels (low / medium / high) so you choose the speed-cost-quality trade-off explicitly. Web grounding and a code interpreter are first-class runtime tools, useful when you want a single model call to actually do the work, not just suggest it.

Best for
  • • Million-token context jobs on a budget — full codebase Q&A, long-form video summarisation, multi-document RAG without per-chunk routing
  • • Agentic workflows that need web grounding or a code interpreter without wiring external tools
  • • AWS-bound enterprise workloads that already use Bedrock — VPC isolation, SageMaker fine-tuning, and AWS-grade SLAs come for free
  • • Multimodal extraction over PDFs and short video clips where Nova Pro is overkill
Reach for something else if
  • • Frontier reasoning or coding benchmarks — route to Claude Opus 4.7 or DeepSeek V3.2 if you need top-of-leaderboard performance
  • • Native audio in/out workflows — Nova 2 Lite does not expose audio as a native modality; use a dedicated speech model
  • • Multi-region routing latency demands — Bedrock pricing carries a ~10% cross-region surcharge when using global endpoints

How TheRouter serves this differently from the vendor

As the vendor operates it

AWS serves Nova 2 Lite through Bedrock bedrock-runtime with Converse, Invoke, response streaming, guardrails, client-side tool calling, prompt caching, extended thinking, web grounding, and code interpreter. The native model card lists text, image, and video input; text output; 1M context; 64K max output; and no native Bedrock structured outputs.

On TheRouter

TheRouter exposes Nova 2 Lite as OpenAI-compatible /v1/chat/completions under amazon/nova-2-lite. Operational routing, pricing, context, modality, and supported-parameter fields are taken from the live TheRouter catalog; this curated page documents AWS-only capabilities and calls out that unproven snippets remain capped until an operator funds verification.

docs.aws.amazon.com/bedrock ↗
Context Length
1M
Max Output
65K
Input Priceper 1M tokens
$0.3564/ 1M tokens
Output Priceper 1M tokens
$2.97/ 1M tokens

Modalities

textimage→text

Capabilities

Vision

Pricing Breakdown

TypeRate
Input$0.3564 / 1M tokens
Output$2.97 / 1M tokens

Supported Parameters

temperaturemax_tokenstop_ptoolstool_choiceresponse_formatreasoningstop

Specifications

Release windowEarly December 2025 (Nova 2 family launch)aws.amazon.com ↗verified
Bedrock model idamazon.nova-2-lite-v1:0 plus us/eu/jp/global inference profile IDsdocs.aws.amazon.com ↗verified
Training cutoffOctober 2025docs.aws.amazon.com ↗verified
Maximum output64,000 tokensdocs.aws.amazon.com ↗verified
Prompt cachingSupported on system and messages fields; 1K minimum per checkpoint, up to 4 checkpoints, 5-minute TTLdocs.aws.amazon.com ↗verified
Extended thinking budget levelsoff (default), low, medium, highaws.amazon.com ↗verified
Built-in toolsWeb grounding (live HTTP with citations), code interpreteraws.amazon.com ↗verified
CustomizationFine-tuning available on Amazon Bedrock and Amazon SageMaker AIaws.amazon.com ↗verified
Bedrock tiersStandard, Priority (+75% premium), Flex (-50% discount); Reserved is not supported for this model cardaws.amazon.com ↗verified
LicenseAWS Service Terms (proprietary, API-only via Bedrock/SageMaker)verified

Benchmarks

BenchmarkDistributionScoreSource
AWS launch benchmark summary
AWS says Nova 2 Lite demonstrates strong performance across instruction following, math, video understanding with temporal reasoning, reliable function calling, precise UI interaction, code generation, and practical software-engineering problem solving. AWS does not expose a complete per-benchmark numeric table on the public launch page, so this page avoids third-party numeric claims until an authoritative source is available.
Qualitative: strong instruction-following, math, video temporal reasoning, function calling, UI interaction, and code generationaws.amazon.com ↗
Long-context benchmark disclosure
Amazon's Nova 2 technical report says it evaluates long-context behavior on LongCodeBench 1M and OpenAI-MRCR, but the public release does not surface Nova 2 Lite's per-model numeric scores. Treat this as an evidence boundary rather than a performance claim.
LongCodeBench 1M and OpenAI-MRCR cited; Nova 2 Lite per-model numbers not publicly disclosedassets.amazon.science ↗
Bedrock model card limits
The current AWS Bedrock model card is the authoritative limits page for operators: it lists model lifecycle Active, context window 1M tokens, maximum output 64K tokens, and knowledge cutoff October 2025.
1M context, 64K max output, Oct 2025 knowledge cutoffdocs.aws.amazon.com ↗

API Usage Examples

Use the global api.therouter.ai endpoint shown below for new integrations; the legacy China accelerated endpoint is retired.

cURL
curl https://api.therouter.ai/v1/chat/completions   -H "Content-Type: application/json"   -H "Authorization: Bearer $THE_ROUTER_API_KEY"   -d '{
    "model": "amazon/nova-2-lite",
    "messages": [
      {"role": "user", "content": "Summarize the key points from this input."}
    ]
  }'

API guide

Full API reference →

Chat completion

Call Nova 2 Lite through TheRouter's OpenAI-compatible surface. TheRouter handles the Bedrock InvokeModel / Converse translation upstream — your client code stays OpenAI-shaped.

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "amazon/nova-2-lite",
    "messages": [{"role": "user", "content": "Summarise this 200-page PDF in 5 bullets."}]
  }'

More from amazon

amazon/nova-proNova flagship — higher accuracy, higher cost; route here when Lite hits its ceiling.amazon/nova-premierTop of the Nova family — for the most demanding agentic and reasoning workloads on Bedrock.amazon/nova-litePrevious-generation Lite — smaller context and no extended-thinking budgets; useful for cost A/B.amazon/nova-microCheapest, text-only Nova — for high-volume classification, routing, and short-form completion.

Similar models

Cross-provider sibling models
anthropic/claude-haiku-4.5
Comparable mini-tier reasoning model on Bedrock with similar price profile; benchmarks reportedly close.
google/gemini-2.5-flash
Google's fast, cheap, long-context option — direct competitor on the cost / context-length axis.
openai/gpt-5-mini
OpenAI's mini tier — different tool surface and pricing structure; useful as the third corner of a routing triangle.
deepseek/deepseek-v3.2
Open-weight alternative for teams that want a self-host fallback alongside Bedrock-hosted Nova.

News & changes

2025-12-02

AWS introduces Nova 2 Lite — 1M context, extended thinking, web grounding built in

AWS announced Nova 2 Lite as the mini-tier of the Nova 2 family, available through Bedrock and SageMaker AI. For TheRouter operators the practical hooks are the 1M context window at mini-tier pricing, opt-in extended thinking budgets (low / medium / high), and first-class web grounding plus code interpreter as runtime tools — useful when one model call needs to do the work, not just describe it.

re-authored by TheRouteraws.amazon.com/blogs/aws ↗
2025-12-02

Amazon publishes Nova 2 technical report

Amazon's Nova 2 technical report covers evaluation methodology (MMLU-Pro, GPQA-Diamond, AIME 2025, IF-Bench, MultiChallenge, LongCodeBench 1M, OpenAI-MRCR) plus safety evaluation under Amazon's Frontier Model Safety Framework. Per-model numeric scores for Nova 2 Lite are sparser than community trackers would like — treat third-party benchmarks (low confidence) accordingly.

re-authored by TheRouterassets.amazon.science ↗

Frequently asked

What is the context window of Nova 2 Lite?

1 million tokens. Large enough for whole-codebase Q&A, long-form video transcripts, or multi-document RAG without chunked routing.

aws.amazon.com ↗
Does Nova 2 Lite support extended reasoning?

Yes, but off by default. Enable via the reasoning field and pick a budget — low, medium, or high — to trade speed and cost for depth. AWS recommends keeping it off until a task actually needs it.

aws.amazon.com ↗
What tools come built in?

Two: web grounding (live HTTP queries with citations) and a code interpreter. Both are managed by Bedrock as runtime tools, separate from any function-calling tools you declare in the request.

aws.amazon.com ↗
Can I fine-tune Nova 2 Lite?

Yes — fine-tuning is offered on both Amazon Bedrock and Amazon SageMaker AI. Useful when prompt engineering hits a ceiling on a narrow vertical workflow.

aws.amazon.com ↗
How does Nova 2 Lite pricing compare across Bedrock tiers?

Bedrock offers three tiers: Standard (baseline), Priority (+75% premium for preferential compute), Flex (-50% discount for workloads that tolerate higher queue time). TheRouter routes Standard-tier traffic by default; ask if you want a Flex-only or Priority-only routing policy.

aws.amazon.com ↗
Fact ledger — every claim on this page traces here
sourceURLretrieved
Release windowaws.amazon.com ↗2026-08-15verified
Bedrock model iddocs.aws.amazon.com ↗2026-08-15verified
Training cutoffdocs.aws.amazon.com ↗2026-08-15verified
Maximum outputdocs.aws.amazon.com ↗2026-08-15verified
Prompt cachingdocs.aws.amazon.com ↗2026-08-15verified
Extended thinking budget levelsaws.amazon.com ↗2026-08-15verified
Built-in toolsaws.amazon.com ↗2026-08-15verified
Customizationaws.amazon.com ↗2026-08-15verified
Bedrock tiersaws.amazon.com ↗2026-08-15verified
License——verified
AWS launch benchmark summaryaws.amazon.com ↗2026-08-15verified
Long-context benchmark disclosureassets.amazon.science ↗2026-08-15verified
Bedrock model card limitsdocs.aws.amazon.com ↗2026-08-15verified
AWS introduces Nova 2 Lite — 1M context, extended thinking, web grounding built inaws.amazon.com/blogs/aws ↗2026-08-15verified
Amazon publishes Nova 2 technical reportassets.amazon.science ↗2026-08-15verified
What is the context window of Nova 2 Lite?aws.amazon.com ↗2026-08-15to verify
Does Nova 2 Lite support extended reasoning?aws.amazon.com ↗2026-08-15to verify
What tools come built in?aws.amazon.com ↗2026-08-15to verify
Can I fine-tune Nova 2 Lite?aws.amazon.com ↗2026-08-15to verify
How does Nova 2 Lite pricing compare across Bedrock tiers?aws.amazon.com ↗2026-08-15to verify
Help & contact