Amazon · AI model
Amazon Nova 2 Lite
Multimodal Amazon Nova model on Amazon Bedrock with a 1M-token context window, up to 65,536 output tokens, optional extended thinking (low/medium/high) and built-in Web Grounding and code interpreter tools. Accepts text, image, video and document input.
Last verified Available 9 sources
Key specs
- Provider
- Amazon
- Family
- Amazon Nova
- API identifier
- amazon.nova-2-lite-v1:0
- Availability
- Available
- Release date
- 2025-12-02
- License
- Proprietary (AWS Customer Agreement)
- Context window
- 1,000,000 tokens
- Max output
- 65,536 tokens
- Knowledge cutoff
- October 2025
- API access
- Yes
- Consumer app
- Unknown
Capabilities
Inputs
- Text input Yes
- Image input Yes PNG, JPEG, GIF, WebP; up to 8,000 x 8,000 px
- Audio input No Bedrock model card lists audio and speech input as not supported; audio tracks in video are not processed
- Video input Yes MP4, MOV, MKV, WebM, FLV, MPEG, MPG, WMV, 3GP; visual frames only
- File input Yes PDF, DOC, DOCX, XLS, XLSX, CSV, TXT, HTML, Markdown; up to 5 documents per request
Outputs
- Text output Yes
- Image output No
- Audio output No
Features
- Reasoning Optional Extended thinking disabled by default; reasoningConfig with maxReasoningEffort low/medium/high
- Tool calling Yes Client-side tool calling; built-in code interpreter per AWS AI Service Card
- Web search Yes Web Grounding system tool (nova_grounding); US Regions and US cross-Region inference profiles only; additional cost
"Unknown" means we have not verified this capability yet; it does not mean the model lacks it.
Pricing
Estimate your monthly cost →| Dimension | Tier | Price | Since | Verified | Source |
|---|---|---|---|---|---|
| Input tokens | base | $0.30 per 1M tokens | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Standard tier, Global cross-Region inference (no in-Region offering in US East). Geo cross-Region (us.) inference: 0.33 USD per 1M tokens. |
| Cached input tokens | base | $0.075 per 1M tokens | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; cache read, Global cross-Region inference. Pricing page: cache read input tokens are 75% less than the on-demand input price (0.075 also listed explicitly for Nova 2 Lite). |
| Output tokens | base | $2.50 per 1M tokens | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Standard tier, Global cross-Region inference; includes reasoning tokens. Geo cross-Region (us.) inference: 2.75 USD per 1M tokens. |
| Batch input tokens | base | $0.15 per 1M tokens | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Batch tier as listed on the pricing page, Global cross-Region inference (Geo: 0.165). |
| Batch output tokens | base | $1.25 per 1M tokens | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Batch tier as listed on the pricing page, Global cross-Region inference (Geo: 1.375). |
| Tool and search requests | base | $30.00 per 1,000 requests | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Web Grounding built-in tool, per 1,000 requests. |
| Other | Flex input | $0.15 per 1M units | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Flex tier, Global cross-Region inference (Geo cross-Region/in-Region: 0.165 USD per 1M tokens). |
| Flex output | $1.25 per 1M units | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Flex tier, Global cross-Region inference (Geo cross-Region/in-Region: 1.375 USD per 1M tokens). | |
| Priority input | $0.525 per 1M units | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Priority tier, Global cross-Region inference (Geo cross-Region/in-Region: 0.5775 USD per 1M tokens). | |
| Priority output | $4.375 per 1M units | 2026-10-12 | 2026-10-25 | Amazon Bedrock Pricing US East (N. Virginia), on-demand list price; Priority tier, Global cross-Region inference (Geo cross-Region/in-Region: 4.8125 USD per 1M tokens). |
Prices are shown in the currency the provider publishes; we never convert currencies.
Price history (10 entries)
| Dimension | Tier | Price | From | Until |
|---|---|---|---|---|
| Input tokens | base | $0.30 per 1M tokens | 2026-10-12 | current |
| Cached input tokens | base | $0.075 per 1M tokens | 2026-10-12 | current |
| Output tokens | base | $2.50 per 1M tokens | 2026-10-12 | current |
| Batch input tokens | base | $0.15 per 1M tokens | 2026-10-12 | current |
| Batch output tokens | base | $1.25 per 1M tokens | 2026-10-12 | current |
| Tool and search requests | base | $30.00 per 1,000 requests | 2026-10-12 | current |
| Other | Flex input | $0.15 per 1M units | 2026-10-12 | current |
| Other | Flex output | $1.25 per 1M units | 2026-10-12 | current |
| Other | Priority input | $0.525 per 1M units | 2026-10-12 | current |
| Other | Priority output | $4.375 per 1M units | 2026-10-12 | current |
Benchmarks
| Benchmark | Setting | Score | Reported by | Measured | Verified | Source |
|---|---|---|---|---|---|---|
| GDPval-AA v2.1 Elo, higher is better | v2.1; high effort; Stirrup agent harness (shell + web browsing) | 366 Elo | Third party | 2026-10-12 | 2026-10-12 | GDPval-AA v2.1 Leaderboard (Artificial Analysis) |
| Text Arena Arena score, higher is better | Overall; style control on (default view); effort not stated | 1,335 | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
Elo is a relative rating: values change as models are added and the scale is re-anchored, so always store the date a rating was read. For example, Anthropic's launch page gives 1846 for Claude Opus 5.5, while the leaderboard on 2026-10-12 shows 1866 (Max, Default Fallback). Earlier versions (Google cites "GDPVal-AA v2") are not comparable. Judging is done by an LLM, not by human experts. Anthropic models are listed with "Default Fallback", meaning refused steps are routed to a fallback model. AA-Briefcase and GDP.pdf are separate evaluations.
Scores measure voter preference, which can be influenced by answer style and length. The leaderboard offers a style-control adjustment, so record which view was used. Scores shift with every update, so a value is only meaningful together with its leaderboard date. Entries marked "Preliminary" or with few votes have wide confidence intervals. The same model can appear several times with different reasoning-effort suffixes (for example -high or -max). Category views (Coding, Math, Creative Writing and others) are different rankings from Overall.
One row per result, without rank numbers: each result was measured under its own setting (effort level, harness, tools, test version), as stated by the source, and scores compare only within the same setting. Self-reported results come from the provider and have not been reproduced independently.
Rankings
- #34 of 42 in Best AI API as of 2026-10-26
- #36 of 39 in Best AI for business as of 2026-10-26
Derived from our published ranking snapshots; each ranking explains its criteria and weights.
Compare Amazon Nova 2 Lite
Guides
- AI token pricing explained: input, output, cached and batch rates
- Reasoning models explained: thinking modes, effort and cost
- Multimodal AI explained: image, audio, video and file inputs, and what they cost
- Context windows explained: limits, long-context pricing and recall
- How to choose an AI model: a practical decision framework
About Amazon Nova 2 Lite
Nova 2 Lite is a multimodal model for text, image, video and document understanding that launched on Amazon Bedrock on December 2, 2025 (model ID amazon.nova-2-lite-v1:0); the Bedrock model card lists an October 2025 knowledge cutoff. Extended thinking is off by default and can be enabled with low, medium or high effort; reasoning tokens are billed as output tokens. The model supports client-side tool calling as well as the built-in Web Grounding and code interpreter tools. In US East (N. Virginia) it is offered through the geo (us.amazon.nova-2-lite-v1:0) and global (global.amazon.nova-2-lite-v1:0) cross-Region inference profiles, not in-Region.
Strengths and weaknesses
Strengths
- 1M-token context window and up to 65,536 output tokens per response
- Optional extended thinking with low, medium and high effort levels
- Built-in Web Grounding (with citations) and code interpreter tools, plus client-side tool calling
- Accepts text, images, video and documents (PDF, DOCX, XLSX, CSV and other formats)
Weaknesses
- Reasoning content is returned as [REDACTED]; reasoning tokens are still billed as output tokens
- Web Grounding is only available in US Regions via US cross-Region inference profiles and is billed separately
- No audio input (audio tracks in video are not processed); structured outputs are not supported according to the Bedrock model card
Sources
- Nova 2 Lite - Amazon Bedrock
- What is Amazon Nova 2? - Amazon Nova
- Amazon Nova 2 Lite - AWS AI Service Cards
- Multimodal understanding - Amazon Nova
- Extended thinking in Amazon Nova 2 - Amazon Nova
- Web Grounding - Amazon Nova
- Amazon Bedrock Pricing
- GDPval-AA v2.1 Leaderboard (Artificial Analysis)
- Text Arena leaderboard (Arena)
Last verified Page updated 2026-10-26