Roboflow

Claude 3 Haiku vs Gemini 3.5 Flash-Lite

Compare Claude 3 Haiku and Gemini 3.5 Flash-Lite side-by-side. See how these vision models stack up in Open Prompt, OCR, Image Captioning, Object Detection, and Classification.

Compare Claude 3 Haiku vs Gemini 3.5 Flash-Lite live

Run the same image across every model that supports a task and compare their outputs side-by-side.

Detect and compare bounding boxes across models on the same image.

Open Object Detection in the full playground
AnthropicClaude 3 Haiku

Claude 3 Haiku is deprecated and can no longer be run. Details and evals are still available on its model page.

GoogleGemini 3.5 Flash-Lite
Run to compare this model.

Models in this comparison

Claude 3 Haiku vs Gemini 3.5 Flash-Lite Comparison Table

Evals updated August 6, 2026Pricing updated August 10, 2026

PropertyClaude 3 HaikuGemini 3.5 Flash-Lite
OrganizationAnthropicGoogle
Categoryclosedclosed
Modalitymultimodalmultimodal
Release DateMar 2024Jul 2026
Context Window200K1.0M
Parameters
LicenseProprietaryProprietary
Pricing per 1M tokens
Input $/1M$0.300
Output $/1M$2.50
Vision Tasks
CaptioningDemo
Chart Question Answering
ClassificationDemo
Document Question Answering
Image Tagging
Multi-Label Classification
Object DetectionDemo
OCRDemo
Vision Language
Visual Question AnsweringDemo
Video Classification
Model Features
Foundation Vision
LLMs with Vision Capabilities
Multimodal Vision
Vision Evalsground-truth scores across 6 vision tasks, pooled at low effort
OverallDeprecated
69.6%
Avg cost / sample$0.0014
Avg speed / sample2.70s
By task
Object Detection
57.5%
$0.0023
Counting
52.7%
$0.0007
Identification
81.3%
$0.0004
OCR
87.4%
$0.0011
Data Extraction
90.7%
$0.0004
Reasoning (low)
48.3%
$0.0012
Reasoning (high)
68.9%
$0.0042

Claude 3 Haiku vs Gemini 3.5 Flash-Lite: Overview

Claude 3 Haiku

Claude 3 Haiku is a large language model developed by Anthropic and released in March 2024 as part of the Claude 3 family, alongside Claude 3 Sonnet and Claude 3 Opus. It is designed to be the fastest and most cost-efficient model in the series, optimized for high-throughput applications.

Like the other Claude 3 models, Haiku is multimodal, able to process both text and image inputs while generating text outputs. It supports a context window of up to 200,000 tokens, with Anthropic noting that the Claude 3 models are technically capable of handling inputs exceeding one million tokens in special cases.

Haiku is positioned as a model well-suited for scenarios that demand speed and scalability at lower cost, such as customer support, summarization, and other tasks where rapid responses are prioritized. Compared to the larger Claude 3 Sonnet and Opus, Haiku provides lower latency and higher efficiency, while the larger models offer stronger reasoning and depth of analysis.

Gemini 3.5 Flash-Lite

Gemini 3.5 Flash-Lite is a natively multimodal reasoning model developed by Google DeepMind, released on July 21, 2026 as part of the Gemini 3.5 model family. It is the fastest model in the 3.5 series, designed for both low-latency tasks and high-throughput production workloads such as agentic search, document processing, receipt translation, and large-scale data extraction. The model accepts text, images, audio, and video as inputs, with a context window of up to 1 million tokens, and produces text output. It supports configurable thinking levels, allowing developers to tune the balance between response quality, cost, and latency depending on workload requirements.

On agentic and coding benchmarks, Gemini 3.5 Flash-Lite significantly outperforms its predecessor, Gemini 3.1 Flash-Lite, including on Terminal-Bench 2.1 (54% vs. 31%), GDM-MRCR v2 long-context (72.2% vs. 60.1%), and real-world task execution as measured by GDPval-AA v2 (1140 vs. 642). It also surpasses Gemini 3 Flash on SWE-Bench Pro (54.2% vs. 49.6%) and OSWorld-Verified (74.0% vs. 65.1%). According to the Artificial Analysis Index, the model generates output at approximately 350 tokens per second. It is built on the Gemini 3.5 Flash foundation and is evaluated across reasoning, coding, multimodal understanding, multilingual performance, and long-context tasks. The model is developed under Google's Frontier Safety Framework.