All AI models · LLM benchmarks · Methodology

Thinking Machines

Inkling

Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio

Model facts

Context window
1.05M tokens
Maximum output
1.05M tokens
Cheapest paid input
$0.95 per 1M tokens
Cheapest paid output
$4.05 per 1M tokens
Open weights
yes
Providers
25

Capabilities and modalities

reasoning, tool calling, structured output, attachments, temperature control, open weights, text, image, audio, pdf.

Local hardware estimate

At 4K context: Q4_K_M 583.9 GiB, Q8_0 1027.4 GiB, F16 1858.9 GiB working memory.

Parameters
952.4 billion
Architecture
inkling
Model license
apache-2.0

Planning estimate adapted from the MIT-licensed whichllm estimator using metadata from Hugging Face; not a vendor minimum requirement.

Artificial Analysis benchmarks

Compare this model on the full LLM benchmark leaderboard.

ConfigurationAAICodingMathOutput tokens/s
xhigh42.352.1not measured70.5

Observed price history

  • 2026-08-24: $0.95 input / $4.05 output per 1M tokens

API providers

  • Kilo Gateway (free) — model id thinkingmachines/inkling:free; free input / free output per 1M tokens; provider documentation
  • Nvidia — model id thinkingmachines/inkling; free input / free output per 1M tokens; provider documentation
  • OpenRouter (free) — model id thinkingmachines/inkling:free; free input / free output per 1M tokens; provider documentation
  • Deep Infra — model id thinkingmachines/Inkling; $0.95 input / $4.05 output per 1M tokens; provider documentation
  • Eden AI — model id deepinfra/thinkingmachines/Inkling; $0.95 input / $4.05 output per 1M tokens; provider documentation
  • Kilo Gateway — model id thinkingmachines/inkling; $0.95 input / $4.05 output per 1M tokens; provider documentation
  • Baseten — model id thinkingmachines/inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • Eden AI — model id together_ai/thinkingmachines/Inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • Fireworks AI — model id accounts/fireworks/models/inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • Hugging Face — model id thinkingmachines/Inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • Merge Gateway — model id thinkingmachines/inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • NanoGPT — model id thinkingmachines/inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • Neon — model id inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • OpenRouter — model id thinkingmachines/inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • OpenRouter (batch) — model id thinkingmachines/inkling:batch; $1 input / $4.05 output per 1M tokens; provider documentation
  • Together AI — model id thinkingmachines/Inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • Vercel AI Gateway — model id thinkingmachines/inkling; $1 input / $4.05 output per 1M tokens; provider documentation
  • Modal — model id thinkingmachines/Inkling-NVFP4; $1.2 input / $5 output per 1M tokens; provider documentation
  • Venice AI — model id inkling; $1.25 input / $5.06 output per 1M tokens; provider documentation
  • Impossibl — model id thinkingmachines/inkling; $1.87 input / $4.68 output per 1M tokens; provider documentation
  • LLMTR — model id thinkingmachines/inkling; $1.87 input / $4.68 output per 1M tokens; provider documentation
  • Requesty — model id inkling; $1.87 input / $4.68 output per 1M tokens; provider documentation
  • Thinking Machines — model id thinkingmachines/Inkling; $1.87 input / $4.68 output per 1M tokens; provider documentation
  • Abacus — model id thinkingmachines/Inkling; $3.74 input / $9.36 output per 1M tokens; provider documentation
  • Arena — model id inkling; not published input / not published output per 1M tokens; provider documentation