Author's Description
Spotlight is a 7‑billion‑parameter vision‑language model derived from Qwen 2.5‑VL and fine‑tuned by Arcee AI for tight image‑text grounding tasks. It offers a 32 k‑token context window, enabling rich multimodal conversations that combine lengthy documents with one or more images. Training emphasized fast inference on consumer GPUs while retaining strong captioning, visual‐question‑answering, and diagram‑analysis accuracy. As a result, Spotlight slots neatly into agent workflows where screenshots, charts or UI mock‑ups need to be interpreted on the fly. Early benchmarks show it matching or out‑scoring larger VLMs such as LLaVA‑1.6 13 B on popular VQA and POPE alignment tests.
Key Specifications
Supported Parameters
This model supports the following parameters:
Performance Summary
Arcee AI's Spotlight, a 7-billion-parameter vision-language model, demonstrates strong performance, particularly in its operational efficiency. It consistently ranks among the fastest models, achieving the 88th percentile in speed across six benchmarks, notably securing the top spot in the Email Classification benchmark and ranking in the top 3 for Coding. The model also offers competitive pricing, typically providing cost-effective solutions at the 74th percentile. Furthermore, Spotlight exhibits exceptional reliability, with a 98th percentile ranking, indicating minimal technical failures and consistent provision of usable responses. In terms of specific benchmark performance, Spotlight shows a mixed but generally positive profile. It excels in Ethics with 99.0% accuracy and strong speed, and demonstrates solid performance in Instruction Following (53.1% accuracy) and General Knowledge (89.1% accuracy), though its percentile rankings in these areas are moderate. A key strength lies in its speed across various tasks, making it highly efficient. However, its accuracy in Coding (71.0%) and Reasoning (50.0%) is in the lower percentiles, suggesting these areas could be considered weaknesses compared to its overall strong speed and reliability. Its design for tight image-text grounding tasks and agent workflows is well-supported by its fast inference capabilities.
Model Pricing
Current Pricing
Feature | Price (per 1M tokens) |
---|---|
Prompt | $0.18 |
Completion | $0.18 |
Price History
Available Endpoints
Provider | Endpoint Name | Context Length | Pricing (Input) | Pricing (Output) |
---|---|---|---|---|
Together
|
Together | arcee-ai/spotlight | 131K | $0.18 / 1M tokens | $0.18 / 1M tokens |
Benchmark Results
Benchmark | Category | Reasoning | Free | Executions | Accuracy | Cost | Duration |
---|
Other Models by arcee-ai
|
Released | Params | Context |
|
Speed | Ability | Cost |
---|---|---|---|---|---|---|---|
Arcee AI: Caller Large Unavailable | May 05, 2025 | — | 32K |
Text input
Text output
|
★★★★ | ★★★ | $$$$ |
Arcee AI: Maestro Reasoning | May 05, 2025 | ~32B | 131K |
Text input
Text output
|
★★ | ★★★★ | $$$$$ |
Arcee AI: Virtuoso Large | May 05, 2025 | ~72B | 131K |
Text input
Text output
|
★★★★★ | ★★★★ | $$$$ |
Arcee AI: Coder Large | May 05, 2025 | ~32B | 32K |
Text input
Text output
|
★★★★★ | ★★★★ | $$$ |
Arcee AI: Virtuoso Medium V2 Unavailable | May 05, 2025 | ~32B | 131K |
Text input
Text output
|
★★★★★ | ★★★★ | $$$ |
Arcee AI: Arcee Blitz Unavailable | May 05, 2025 | ~24B | 32K |
Text input
Text output
|
★★★★ | ★★★★ | $$$ |