LLM Pricing
ModelsLeaderboardsToolsProvidersReleasesArticles

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

Inkling 2 Small

thinkingmachines·thinkingmachines/inkling-2-small·GA·Open weights·ling series·NEW
Multimodal MoE reasoning model (594B total, 17B active) for agentic coding, vision, and knowledge work

Specs & pricing

Input / output per 1M tokens
Official price
—
Lowest paid·Kilo GatewayGateway
$0.15 / $0.50
Blended $0.24 · 2× spread
2 more $0 channels
Context
1,048,000
Output limit
131,072
Knowledge cutoff
—
Released / updated
2026-11-05 / 2026-11-12
Capabilities
✓ Reasoning✓ Tool use✓ Structured output✓ Temperature✓ Attachments
Modalities
TextImage
License
Apache-2.0
Weights
Hugging Face (NVFP4)nvfp4Hugging Face (MXFP8)mxfp8
References
announcementmodel card

Inkling 2 Small by thinkingmachines is offered by 18 providers on this page. Public prices are shown for 16 of them. No official API price is listed for it; the prices below come from third-party channels. The lowest paid channel is Kilo Gateway at $0.15 / $0.50 per 1M. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. Prices vary about 2.0× across the providers that host it. It also has 2 free ($0) channels; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.

Artificial Analysis rates it 30.6 on the Intelligence Index, with 44.3 for Finance & Accounting, its strongest domain. Against its lowest blended price of $0.237 per 1M, that is roughly 129 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 112 tokens per second, with 2.37s to the first token. Running one task of the Artificial Analysis suite costs about $1.1784, which reflects how many tokens its reasoning consumes rather than the unit price alone.

The context window is 1,048,000 tokens at the reference host, but hosts report different limits, from 524,288 to 1,048,576, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text and Image. The weights are open, so it can also be self-hosted or served through a gateway of your choice.

Quality & performanceArtificial Analysis · Intelligence Index v5

Intelligence30.6
Value129
Engineering38.9
Finance & Accounting44.3
Legal41.6
Healthcare & Medical40.8
Strategy & Ops43.5
Economics43.4
Output speed112 tok/s
TTFT2.37 s
Cost per task$1.18
Value formulaIQ 30.6 ÷ min blended $0.237 = 129

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

Available at 18 providers16 with public prices · 2 free

  • ProviderInput / Output
  • OpenRouterGateway
    thinkingmachines/inkling-2-small:free
    Free
    Context 1,048,576 ⚠Output limit 131,072
  • OpenCodeGateway
    inkling-2-small-free
    Free
    Cache read $0Context 1,048,576 ⚠Output limit 131,072
  • Kilo GatewayGateway
    $0.15 / $0.50
    Cache read $0.03Context 1,048,576 ⚠Output limit 943,718
  • Prime InferenceGateway
    $0.15 / $0.50
    Cache read $0.03Context 1,048,576 ⚠Output limit 1,048,576
  • Deep InfraCloud
    thinkingmachines/Inkling-2-Small
    $0.30 / $1.00
    Cache read $0.03Context 1,048,576 ⚠Output limit 131,072
  • OpenRouterGateway
    $0.30 / $1.00
    Cache read $0.03Context 1,048,576 ⚠Output limit 943,718
  • Vercel AI GatewayCloud
    $0.30 / $1.00
    Cache read $0.03Context 1,048,576 ⚠Output limit 262,144
  • Ollama CloudCloud
    $0.30 / $1.00
    Cache read $0.03Context 1,048,576 ⚠Output limit 131,072
  • NanoGPTGateway
    $0.30 / $1.00
    Cache read $0.03Context 524,288 ⚠Output limit 131,072
  • Nebius Token FactoryCloud
    thinkingmachines/Inkling-2-Small
    $0.30 / $1.00
    Cache read $0.06Context 1,048,576 ⚠Output limit 1,048,576
ProviderTierInputOutputCache readCache writeContextOutput limitStatus
OpenRouter
thinkingmachines/inkling-2-small:free
GatewayFree——1,048,576 ⚠131,072
OpenCode
inkling-2-small-free
GatewayFree$0$01,048,576 ⚠131,072
Kilo GatewayGateway$0.15$0.50$0.03—1,048,576 ⚠943,718
Prime InferenceGateway$0.15$0.50$0.03$0.151,048,576 ⚠1,048,576
Deep Infra
thinkingmachines/Inkling-2-Small
Cloud$0.30$1.00$0.03—1,048,576 ⚠131,072
OpenRouterGateway$0.30$1.00$0.03—1,048,576 ⚠943,718
Vercel AI GatewayCloud$0.30$1.00$0.03—1,048,576 ⚠262,144
Ollama CloudCloud$0.30$1.00$0.03—1,048,576 ⚠131,072
NanoGPTGateway$0.30$1.00$0.03—524,288 ⚠131,072
Nebius Token Factory
thinkingmachines/Inkling-2-Small
Cloud$0.30$1.00$0.06—1,048,576 ⚠1,048,576

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

effort = minimaleffort = loweffort = mediumeffort = higheffort = maxeffort = noneeffort = xhighToggle (on / off)

Interleaved thinking (reasoning between tool calls) is declared by 2 of 18 providers.

Your usage cost

Common scenarios
1Kilo Gateway$40.60
2Prime Inference$40.60
3Deep Infra$77.60
4OpenRouter$77.60
5Vercel AI Gateway$77.60
6Ollama Cloud$77.60

2 more channels offer $0 (OpenRouter, OpenCode); free tiers usually have rate limits and no SLA, excluded from ranking.

The cheapest paid channel is the only channel.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$0.24$02026-11-072026-11-13
2026-11-13Input list —Output list —Min blended $0.24

Benchmark1 items

NameConditionsScoreMetricSource
Artificial Analysis Intelligence Indexvv5.030.6—Source ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Artificial Analysis evaluations24 items

Humanity's Last Exam44.1%
HLE-Diamond28.7%
MMMU-Pro81.6%
CritPt v2 Hard35.0%
SciCode52.3%
FrontierCode 2.025.4%
Terminal-Bench 2.181.3%
Terminal-Bench 4.013.1%
Terminal-Bench-Science5.7%
τ³-Bench Banking40.4%
AutomationBench-AA45.6%
AA-AnalystAgent51.2%
EnterpriseOps-Gym-AA46.8%
AA-Briefcase1527
GDPval-AA1630
GDP.pdf25.2%
AA-LCR83.3%
MLCR-AA20.6%
AA-Omniscience Index16.1
AA-Omniscience Accuracy33.8%
AA-Omniscience Non-Hallucination73.3%
AA-3D1027
Code2Video-AA1020
vials-aa0.4286

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

inLing 3.2 Flash (free)same series—inLing 3.2 Flashsame series—ininclusionAI: Ling 3.1 Flash 1019same series—Inklingsame lab$1.87 / $4.68Inkling Smallsame lab$0.58 / $1.44
Data partly from models.dev (MIT)