---
id: "thinkingmachines/inkling"
also: ["thinkingmachines/inkling-20260715", "thinkingmachines/Inkling", "hf/thinkingmachines/Inkling", "requesty/thinkingmachines/inkling", "vercel/thinkingmachines/inkling"]
cheapest_list_usd_per_m: "0.95 in, 4.05 out, at deepinfra on hf"
cheapest_effective_usd_per_m: "0.5926 in, 4.05 out, at thinkingmachines on vercel"
context: 524288
max_output: 471859
input: ["text", "image", "audio"]
output: ["text"]
released: "2026-07-17"
reasoning: ["max", "high", "medium", "low", "minimal", "none"]
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "include_reasoning", "logit_bias", "max_tokens", "min_p", "presence_penalty", "reasoning", "reasoning_effort", "repetition_penalty", "response_format", "seed", "stop", "structured_outputs", "temperature", "tool_choice", "tools", "top_k", "top_p"]
endpoints: 13
gateways: "vercel (5), hf (4), openrouter (3), requesty (1)"
no_training: 8
trains: 1
retains_prompts: 2
intelligence: 12.5000
coding: 52.1000
agentic: 22.5000
---
# Thinking Machines: Inkling

Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems,...