LLM Pricing
ModelsLeaderboardsToolsProvidersReleasesArticles

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

Inkling Small

thinkingmachines·thinkingmachines/inkling-small·GA·Open weights·ling series
Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio

Specs & pricing

Input / output per 1M tokens
Official price·Thinking Machines
$0.58 / $1.44
Blended $0.80 · Cache read $0.12
Lowest paid·Thinking MachinesFirst-party
$0.30 / $1.20
Blended $0.53 · 2× spread
Context
65,536
Output limit
65,536
Knowledge cutoff
—
Released / updated
2026-07-30 / 2026-07-30
Capabilities
✓ Reasoning✓ Tool useStructured output✓ Temperature✓ Attachments
⚠ Providers report capability flags inconsistently
Modalities
TextImageAudioPDF
License
Apache-2.0
Weights
Hugging Face

Inkling Small by thinkingmachines is offered by 15 providers on this page. Its official list price is $0.58 per 1M input tokens and $1.44 per 1M output tokens. The lowest paid channel is Thinking Machines at $0.30 / $1.20 per 1M, about 1.5× below the list price.

Artificial Analysis rates it 12.9 on the Intelligence Index, with 24.9 for Economics, its strongest domain. Against its lowest blended price of $0.525 per 1M, that is roughly 25 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 56 tokens per second, with 2.21s to the first token. Running one task of the Artificial Analysis suite costs about $0.2032, which reflects how many tokens its reasoning consumes rather than the unit price alone.

The context window is 65,536 tokens at the reference host, but hosts report different limits, from 65,536 to 1,000,000, so the usable window depends on the provider you pick. It supports reasoning and tool use. Accepted input modalities are Text, Image, Audio, and PDF. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use.

Quality & performanceArtificial Analysis · Intelligence Index v5

Intelligence12.9
Value25
Engineering23.1
Finance & Accounting23.7
Legal21.7
Healthcare & Medical18.6
Strategy & Ops18
Economics24.9
Output speed56 tok/s
TTFT2.21 s
Cost per task$0.20
Value formulaIQ 12.9 ÷ min blended $0.525 = 25
Reasoning tier → intelligence / speed (higher tier = stronger but slower)
HighIQ 9.6 · 60 tok/s
defaultIQ 12.9 · 56 tok/s

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 15 providers15 with public prices

  • ProviderInput / Output
  • Thinking MachinesOfficialFirst-party
    thinkingmachines/Inkling-Small:peft:262144:sampling-nvfp4
    $0.30 / $1.20
    Cache read $0.06Context 262,144 ⚠Output limit 262,144
  • NanoGPTGateway
    thinkingmachines/Inkling-Small
    $0.45 / $1.20
    Cache read $0.10Context 524,288 ⚠Output limit 32,768
  • NanoGPTGateway
    thinkingmachines/Inkling-Small:thinking
    $0.45 / $1.20
    Cache read $0.10Context 524,288 ⚠Output limit 32,768
  • OpenRouterGateway
    $0.45 / $1.20
    Cache read $0.10Context 524,288 ⚠Output limit 262,144
  • Eden AIGateway
    deepinfra/thinkingmachines/Inkling-Small
    $0.45 / $1.20
    Cache read $0.10Context 524,288 ⚠Output limit 524,288
  • Kilo GatewayGateway
    $0.45 / $1.20
    Cache read $0.10Context 524,288 ⚠Output limit 262,144
  • Vercel AI GatewayCloud
    $0.45 / $1.20
    Cache read $0.10Context 1,000,000 ⚠Output limit 1,000,000
  • Deep InfraCloud
    thinkingmachines/Inkling-Small
    $0.45 / $1.20
    Cache read $0.10Context 524,288 ⚠Output limit 524,288
  • ArceeFirst-party
    $0.50 / $1.20
    Cache read $0.10Context 262,144 ⚠Output limit 262,144deprecated
  • CommandCodeGateway
    $0.50 / $1.20
    Cache read $0.10Context 1,000,000 ⚠Output limit 1,000,000
ProviderTierInputOutputCache readCache writeContextOutput limitStatus
Thinking MachinesOfficial
thinkingmachines/Inkling-Small:peft:262144:sampling-nvfp4
First-party$0.30$1.20$0.06—262,144 ⚠262,144
NanoGPT
thinkingmachines/Inkling-Small
Gateway$0.45$1.20$0.10—524,288 ⚠32,768
NanoGPT
thinkingmachines/Inkling-Small:thinking
Gateway$0.45$1.20$0.10—524,288 ⚠32,768
OpenRouterGateway$0.45$1.20$0.10—524,288 ⚠262,144
Eden AI
deepinfra/thinkingmachines/Inkling-Small
Gateway$0.45$1.20$0.10—524,288 ⚠524,288
Kilo GatewayGateway$0.45$1.20$0.10—524,288 ⚠262,144
Vercel AI GatewayCloud$0.45$1.20$0.10—1,000,000 ⚠1,000,000
Deep Infra
thinkingmachines/Inkling-Small
Cloud$0.45$1.20$0.10—524,288 ⚠524,288
ArceeFirst-party$0.50$1.20$0.10—262,144 ⚠262,144deprecated
CommandCodeGateway$0.50$1.20$0.10—1,000,000 ⚠1,000,000

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Reasoning control

Toggle (on / off)effort = loweffort = mediumeffort = higheffort = xhigheffort = maxeffort = minimaleffort = none

Interleaved thinking (reasoning between tool calls) is declared by 2 of 15 providers.

3 / 15 providers expose no reasoning control (reasoning_options: []).

Your usage cost

Common scenarios
1Thinking Machines · Official$91.20
2NanoGPT$108.00
3NanoGPT$108.00
4OpenRouter$108.00
5Eden AI$108.00
6Kilo Gateway$108.00
The cheapest paid channel is official.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$1.44$02026-08-052026-10-21
2026-10-21Input list $0.58Output list $1.44Min blended $0.53

Artificial Analysis evaluations25 items

GPQA Diamond89.5%
Humanity's Last Exam33.3%
HLE-Diamond12.1%
MMMU-Pro74.1%
CritPt8.3%
CritPt v2 Hard18.5%
SciCode49.6%
FrontierCode0.5%
FrontierCode 2.04.8%
Terminal-Bench 2.155.1%
Terminal-Bench 4.01.0%
Terminal-Bench-Science0.0%
τ³-Bench Banking18.8%
AutomationBench-AA4.8%
AA-AnalystAgent27.5%
EnterpriseOps-Gym-AA42.7%
AA-Briefcase913
GDPval-AA1133
GDP.pdf11.2%
AA-LCR75.7%
MLCR-AA7.8%
AA-Omniscience Index-9.5
AA-Omniscience Accuracy31.8%
AA-Omniscience Non-Hallucination39.5%
Harvey LAB-AA0.0%

Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and scale (percent correct, points or an index), so rows are not comparable with one another.

Related models

ininclusionAI: Ling 3.2 Flash (free)same series—Inkling 2 Smallsame series—inLing 3.2 Flashsame series—GLM-5.3-Flashcheaper alternative$0.15 / $0.50DeepSeek V4.1 Flash 1017cheaper alternative$0.15 / $0.60GPT-5.6 Lunacheaper alternative$0.20 / $1.20Claude Haiku 5.5cheaper alternative$0.10 / $0.50Inklingsame lab$1.87 / $4.68
Data partly from models.dev (MIT) · Thinking Machines official docs ↗