Nova Lite vs Phi-4-multimodal-instruct vs Gemini 2.0 Flash-Lite
Gemini 2.0 Flash-Lite comes out ahead, 82 to 75 and 72 on our weighted score, though Phi-4-multimodal-instruct is 28% cheaper per token.
AmazonNova Lite
75/100- ECI—
- Price$0.06 / $0.24
- Context300K
MicrosoftPhi-4-multimodal-instruct
72/100- ECI—
- Price$0.07 / $0.11
- Context131K
- Our pick
GoogleGemini 2.0 Flash-Lite
82/100- ECI—
- Price$0.064 / $0.255
- Context1.05M
Gemini 2.0 Flash-Lite is our pick
Gemini 2.0 Flash-Lite is the better all-round choice, scoring 82/100 against Nova Lite (75) and Phi-4-multimodal-instruct (72). It leads on inputs & features and context window. Phi-4-multimodal-instruct wins on price. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
- CapabilityNot enough dataNo independent benchmark covers every model here yet
- Lowest pricePhi-4-multimodal-instructPhi-4-multimodal-instruct $0.08 · Nova Lite $0.105 · Gemini 2.0 Flash-Lite $0.111 per 1M tokens (3:1 blend)
- Longest contextGemini 2.0 Flash-LiteGemini 2.0 Flash-Lite 1,048,576 · Nova Lite 300,000 · Phi-4-multimodal-instruct 131,072 tokens
- Widest inputsGemini 2.0 Flash-LiteNova Lite: Text, Images, PDFs, Video · Phi-4-multimodal-instruct: Text, Images, Audio · Gemini 2.0 Flash-Lite: Text, Images, PDFs, Audio, Video
- Self-hostingPhi-4-multimodal-instructPublishes downloadable weights (MIT)
| Measure | Weight | Nova Lite | Phi-4-multimodal-instruct | Gemini 2.0 Flash-Lite |
|---|---|---|---|---|
| Price | 50% | 92 | 98 | 91 |
| Inputs & features | 30% | 70 | 60 | 80 |
| Context window | 20% | 39 | 25 | 61 |
| Overall | 100% | 75/100 | 72/100 | 82/100 |
Left out because at least one model lacks the data: capability. The remaining weights were rescaled.
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | Nova LiteAmazon | Phi-4-multimodal-instructMicrosoft | Gemini 2.0 Flash-LiteGoogle |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | $0.06 (best) | $0.07 | $0.064 |
| Output | $0.24 | $0.11 (best) | $0.255 |
| Cached input | $0.015 | — | — |
| Blended (3:1) | $0.105 | $0.08 (best) | $0.111 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Official Amazon Bedrock API | Median of 1 providers | Median of 2 providers |
| Limits | |||
| Context window | 300,000 tokens | 131,072 tokens | 1,048,576 tokens (best) |
| Max output | 10,000 tokens | 131,072 tokens (best) | 8,192 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | Yes | No | Yes |
| Audio | No | Yes | Yes |
| Video | Yes | No | Yes |
| Reasoning | No | No | No |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | No | No |
| Availability | |||
| Weights | Proprietary | OpenMIT | Proprietary |
| API model ID | amazon.nova-lite-v1:0 | — | — |
| API providers | 8 (best) | 1 | 3 |
| Released | Dec 3, 2024 | Feb 26, 2025 | Dec 11, 2024 |
| Knowledge cutoff | Oct 2024 | Jun 2024 | Jun 2024 |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Nova Lite$1.08
- Phi-4-multimodal-instruct$0.92
- Gemini 2.0 Flash-Lite$1.15
Which should you choose?
Which is better: Nova Lite, Phi-4-multimodal-instruct or Gemini 2.0 Flash-Lite?
Gemini 2.0 Flash-Lite is the better all-round choice, scoring 82/100 against Nova Lite (75) and Phi-4-multimodal-instruct (72). It leads on inputs & features and context window. Phi-4-multimodal-instruct wins on price. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
Which is cheaper, Nova Lite, Phi-4-multimodal-instruct or Gemini 2.0 Flash-Lite?
Phi-4-multimodal-instruct is cheaper at $0.07 input / $0.11 output per million tokens (median across 1 API provider). Nova Lite costs $0.06 input / $0.24 output per million tokens (official Amazon Bedrock API price); Gemini 2.0 Flash-Lite costs $0.064 input / $0.255 output per million tokens (median across 2 API providers). At a typical mix of three input tokens to one output token, that is $0.08 per million tokens for Phi-4-multimodal-instruct versus $0.105 for Nova Lite (1.3× as much) and $0.111 for Gemini 2.0 Flash-Lite (1.4× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. Nova Lite has not been scored yet, Phi-4-multimodal-instruct has not been scored yet and Gemini 2.0 Flash-Lite has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for Nova Lite, Phi-4-multimodal-instruct and Gemini 2.0 Flash-Lite yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. All three support tool calling for agent workflows.
Which has the bigger context window?
Gemini 2.0 Flash-Lite has the largest context window at 1,048,576 tokens, against 300,000 for Nova Lite and 131,072 for Phi-4-multimodal-instruct. Maximum output per response: Nova Lite up to 10,000, Phi-4-multimodal-instruct up to 131,072, Gemini 2.0 Flash-Lite up to 8,192 tokens.
Which can read images, PDFs, audio or video?
Nova Lite accepts text, images, PDFs and video; Phi-4-multimodal-instruct accepts text, images and audio; Gemini 2.0 Flash-Lite accepts text, images, PDFs, audio and video. Gemini 2.0 Flash-Lite handles the widest range of inputs.
Are any of these open source?
Phi-4-multimodal-instruct publishes its weights (MIT) and can be self-hosted; Nova Lite and Gemini 2.0 Flash-Lite is proprietary.
Which is newer?
Phi-4-multimodal-instruct is the newest, released Feb 26, 2025. Gemini 2.0 Flash-Lite came out Dec 11, 2024; Nova Lite came out Dec 3, 2024. Knowledge cutoff: Nova Lite Oct 2024, Phi-4-multimodal-instruct Jun 2024, Gemini 2.0 Flash-Lite Jun 2024.
How do you decide the winner?
Each model gets a 0–100 score on capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 2, 2026.
