togetherai/capabilities

6 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

Together.ai reasoning models with extractReasoningMiddleware

Together.ai exposes the thinking of `deepseek-ai/DeepSeek-R1` in generated text using `<think>` tags. The `extractReasoningMiddleware` from the `ai` package can be used with `wrapLanguageModel` to extract this reasoning and expose it as a `reasoning` property on the result.

Together.ai language model capabilities table

Language model capabilities by model: moonshotai/Kimi-K2.5 (Image Input: yes, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes); Qwen/Qwen3.5-397B-A17B (Image Input: yes, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes); MiniMaxAI/MiniMax-M2.5 (Image Input: no, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes); zai-org/GLM-5 (Image Input: no, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes); deepseek-ai/DeepSeek-V3.1 (Image Input: no, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes); openai/gpt-oss-120b (Image Input: no, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes); openai/gpt-oss-20b (Image Input: no, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes); meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 (Image Input: yes, Object Generation: yes, Tool Usage: yes, Tool Streaming: yes).

Together.ai image editing with FLUX Kontext models

Together AI supports image editing through FLUX Kontext models by passing input images via `prompt.images` to transform or edit existing images. Together AI does not support mask-based inpainting and instead uses descriptive prompts to specify changes. Input images can be provided as Buffer, ArrayBuffer, Uint8Array, base64-encoded strings, or URLs, with only a single input image supported per request.

Together.ai image model size support

Together.ai image models support various image dimensions that vary by model. Common sizes include 512x512, 768x768, and 1024x1024, with some models supporting up to 1792x1792. The default size is 1024x1024.

Together.ai embedding model capabilities

Embedding model capabilities: BAAI/bge-large-en-v1.5 (Dimensions: 1024, Max Tokens: 512); Alibaba-NLP/gte-modernbert-base (Dimensions: 768, Max Tokens: 8192); intfloat/multilingual-e5-large-instruct (Dimensions: 1024, Max Tokens: 514).

Together AI models support matrix

Together AI models and their capabilities: meta-llama/Meta-Llama-3.3-70B-Instruct-Turbo does not support Image Input, Object Generation, Tool Usage, or Tool Streaming; Qwen/Qwen2.5-72B-Instruct-Turbo does not support Image Input, Object Generation, Tool Usage, or Tool Streaming; deepseek-ai/DeepSeek-V3 does not support Image Input, Object Generation, Tool Usage, or Tool Streaming; mistralai/Mixtral-8x22B-Instruct-v0.1 does not support Image Input but supports Object Generation, Tool Usage, and Tool Streaming.

Give your agent this brain