Reka Models by Reka
Details
Capabilities
About
Reka AI offers an impressive suite of multimodal large language models, adept at processing various data forms such as text, images, video, and audio. The lineup includes models like Reka Core, Flash, Edge, and Spark, each tailored for different computational needs and deployment contexts. Reka Core stands out as a "frontier-class" model, often compared with top-tier models from OpenAI, Google, and Anthropic, achieving competitive results on numerous benchmarks. Flash and Edge provide robust performance in their compute categories, frequently outpacing larger models. Built on a novel multimodal architecture, these models support advanced capabilities like multimodal understanding, reasoning, code generation, and multilingual processing. They are available for use via API, on-premise, or on-device, ensuring flexible deployment options.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
Use when the workload needs 8k context, 21B parameters, and multimodal inputs.
Use when the workload needs 64k context, 21B parameters, and tool use.
Use when the workload needs 128k context, function calling, and multimodal inputs.
Use when the workload needs 128k context, 21B parameters, and multimodal inputs.
Use when the workload needs 64k context, 7B parameters, and structured outputs.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Reka Flash 3.1 | Use when the workload needs 8k context, 21B parameters, and multimodal inputs. | 2026-03 | 8k context21B parametersmultimodal inputs | Current |
| Reka Flash 3 | Use when the workload needs 64k context, 21B parameters, and tool use. | 2025-12 | 64k context21B parameterstool use | Current |
| Reka Core | Use when the workload needs 128k context, function calling, and multimodal inputs. | 2024-04 | 128k contextfunction callingmultimodal inputs | Current |
| Reka Flash | Use when the workload needs 128k context, 21B parameters, and multimodal inputs. | 2024-02 | 128k context21B parametersmultimodal inputs | Current |
| Reka Edge | Use when the workload needs 64k context, 7B parameters, and structured outputs. | 2024-02 | 64k context7B parametersstructured outputs | Current |
Release Timeline
4 release groupsSpecifications(5 models)
| Model | Released | Context | Parameters | Multimodal | Fn Calling | Tool Use | Structured Outputs |
|---|---|---|---|---|---|---|---|
| Reka Flash 3.1 | 2026-03 | 8k | 21B | Yes | No | No | No |
| Reka Flash 3 | 2025-12 | 64k | 21B | No | Yes | Yes | Yes |
| Reka Core | 2024-04 | 128k | — | Yes | Yes | No | No |
| Reka Flash | 2024-02 | 128k | 21B | Yes | No | No | No |
| Reka Edge | 2024-02 | 64k | 7B | Yes | No | No | Yes |
Available From(2 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Reka Edge | Reka Platform | $0.1 | $0.4 | Serverless |
| Reka Edge | OpenRouter | $0.1 | $0.1 | Serverless |
| Reka Flash | Reka Platform | $0.2 | $0.8 | Serverless |
| Reka Core | Reka Platform | $2 | $6 | Serverless |
Frequently Asked Questions
- What is Reka used for?
- Reka is used for vision and multimodal work, agent workflows and tool use, and structured outputs. The family description and listed model capabilities point to those workloads as the best fit.
- How does Reka compare to Claude 4.8?
- Reka by Reka is strongest where you need vision and multimodal work, while Claude 4.8 by Anthropic is the closest related family to check for vision and multimodal work. Reka has 5 listed variants and reaches up to 128k context, while Claude 4.8 reaches up to 1m context, so compare the specs and pricing tables before choosing a production model.
- Which Reka model should I use?
- For the lowest listed input price, start with Reka Edge through Reka Platform at $0.1/1M input tokens. For the most capable/latest local choice, evaluate Reka Flash 3 with 64k context and tool use, function calling, and structured outputs.
