Thinking Machines

Inkling

CTX66K

77Quality

Reasoning · Vision · Tools

Input$1.87/M

Output$4.68/M

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

Open weightsOpen license949/1040

Drag to turn

Thinking MachinesTextUncommon

Inkling

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

thinkingmachines/Inkling

Type: Text
Quality: 77 · Top 23% on the Epoch Capabilities Index
Context window: 65,536 tokens
Max output: 65,536 tokens
Input price: $1.87 / 1M tokens
Output price: $4.68 / 1M tokens
Cache read: $0.37 / 1M tokens
Cache write: —
Released: Jul 15, 2026
Knowledge cutoff: —
Takes in: Text, Image
Puts out: Text
Reasoning: Yes
Tool calling: Yes
Structured output: No
File attachments: Yes
Access: Open weights
License: Open license
Rarity: Uncommon
Set number: 949/1040

Open Thinking Machines docs Quality 77 Where to run it 25 Meet its creature Battle QVQ Max

No matching Hugging Face repo was found, so this shows Thinking Machines’s usual open-weight license. Check the model card before relying on it. Specs from models.dev, last changed Oct 18, 2026. Official source: tinker-docs.thinkingmachines.ai.

How good is it?

Public leaderboard scores for Inkling, refreshed daily from Epoch AI and LMArena.

Top 23% of the 277 models on the Epoch Capabilities Index: #65.

Quality is the share of a leaderboard’s models this one beats, from 0 to 100. Chat models are rated on Epoch AI’s Capabilities Index when it lists them, otherwise on LMArena; image and video models on LMArena’s image and video arenas.

Leaderboards

Epoch Capabilities Index#65 of 277
One number fitted from dozens of hard benchmarks, from science questions to agentic coding.148.6 index
LMArena Text#74 of 416
People compare two anonymous answers and vote for the better one.1,441 rating
LMArena Text · Coding#78 of 411
The text arena’s coding prompts only.1,465 rating
LMArena Text · Math#34 of 400
The text arena’s math prompts only.1,477 rating
LMArena Text · Creative writing#121 of 414
The text arena’s creative writing prompts only.1,387 rating
LMArena WebDev#87 of 144
Each model builds a web app from the same prompt; people pick the better one.1,414 rating

Scores from Epoch AI and LMArena, licensed CC BY 4.0. Boards list models under their own names, so a close match is shown with the name it was scored under. Credits

Where to run it

25 providers sell Inkling through an API. Prices are per million tokens, cheapest first.

CheapestDeep Infra$0.95 in · $4.05 out
Free to try2 providersNvidia, OpenRouter
From the labThinking Machines$1.87 in · $4.68 out

API providers

Provider	Input	Output	Context
NvidiaFree`thinkingmachines/inkling`	Free	Free	1.05M
OpenRouterFree`thinkingmachines/inkling:free`	Free	Free	1.05M
Deep InfraCheapest paid`thinkingmachines/Inkling`	$0.95	$4.05	524K
DevPass (LLM Gateway)`inkling`	$0.95	$4.05	524K
Eden AI`deepinfra/thinkingmachines/Inkling`	$0.95	$4.05	524K
Kilo Gateway`thinkingmachines/inkling`	$0.95	$4.05	524K
LLM Gateway`deepinfra/inkling`	$0.95	$4.05	524K
NanoGPT`thinkingmachines/inkling`	$0.95	$4.05	1.05M
Baseten`thinkingmachines/inkling`	$1.00	$4.05	1.05M
CommandCode`thinkingmachines/inkling`	$1.00	$4.05	256K

Show 15 more providers

Provider	Input	Output	Context
Fireworks AI`accounts/fireworks/models/inkling`	$1.00	$4.05	1.05M
Hugging Face`thinkingmachines/Inkling`	$1.00	$4.05	1.05M
Merge Gateway`thinkingmachines/inkling`	$1.00	$4.05	1.05M
Neon`inkling`	$1.00	$4.05	1.05M
Together AI`thinkingmachines/Inkling`	$1.00	$4.05	524K
Vercel AI Gateway`thinkingmachines/inkling`	$1.00	$4.05	256K
Charm Hyper`inkling`	$1.09	$4.41	1.05M
Impossibl`thinkingmachines/inkling`	$1.87	$4.68	66K
LLMTR`thinkingmachines/inkling`	$1.87	$4.68	262K
Requesty`inkling`	$1.87	$4.68	66K
Thinking MachinesOfficial`thinkingmachines/Inkling`	$1.87	$4.68	66K
Modal`thinkingmachines/Inkling-NVFP4`	$1.20	$5.00	1.05M
Venice AI`inkling`	$1.25	$5.06	524K
AI/ML API`thinkingmachines/inkling`	$1.38	$5.57	1.05M
Abacus`thinkingmachines/Inkling`	$3.74	$9.36	262K

From models.dev and Hugging Face, checked daily. Resellers’ listings can lag behind their price pages, so confirm with the provider before you commit.

Face off

Current text models from other labs at a similar output price.