Skip to content
ModelCap

Inkling 2 Small

Arena #152Official publisher releaseOpen weightsFree tier6.3d #95 → #98

Thinking Machines·#98 of 371 on the current board·cheaper than #96 Mistral Large 4

Inkling 2 Small is a language model from Thinking Machines. It ranks #98 of 371 on ModelCap's current board, with an Index score of 56.8, measured on 5 public benchmarks. Its score range of 50.7–62.9 spans positions #70–#135. It is listed at $0.30 per million input tokens and $1.00 per million output tokens across 6 providers. It reads up to 1M tokens of context, and its weights are openly downloadable.

Figures as of 19 Nov 2026, 05:16 UTC · download the public dataset · methodology · launch report

ModelCap Index · Current board
56.8#98
Measured on 5 public benchmarks · range 50.7–62.9
Evidence as of 18 Nov 2026, 14:36 UTC
Market Gravity
53.8Decreased by 0.2
Secondary market signal
Input
$0.30
per 1M tokens
Output
$1.00
per 1M tokens
Context
1M
tokens
Providers
6
endpoints

Providers

Uptime measured over the last 30 minutes
ProviderInputOutputContextQuantUptime
PrimeIntellectfp4Cheapest$0.15$0.501Mfp4100.0%
DeepInfrafp4$0.285$0.951Mfp460.00%
Cloudflare$0.30$1.001M—100.0%
Modalnvfp4$0.30$1.001Mnvfp4100.0%
Nebiusfp4$0.30$1.001Mfp4100.0%
Together$0.30$1.001M—100.0%

Hugging Face Inference Providers

Exact repository identity

Provider economics and telemetry published by the Hugging Face Router. These are operational signals, not benchmark results.

ProviderInputOutputContextTTFTThroughputFeatures
primeintellect
live
$0.15$0.501M764 ms118.0 tok/stools · structured
cloudflare
live
$0.30$1.001M1473 ms99.0 tok/stools · structured
deepinfra
live
$0.30$1.001M462 ms140.5 tok/stools · structured
modal
live
$0.30$1.001M559 ms200.0 tok/stools · structured
nebius
live
$0.30$1.001M2495 ms64.0 tok/stools · structured
together
live
$0.30$1.001M390 ms208.0 tok/stools · structured

Reasoning efforts

Best supported result can inform ModelCap

Published Max, X-High, High and other configurations for this model. These are sourced scores, not head-to-head essays.

EffortAgent scoreSource rankSampleSource
Default
Inkling 2 Small
1.51%
CI 0.5–2.5%
#263.5M obs
22K sessions
LMArena Agent
Match 100%

Reported evaluation leads (15)

Not independent measurement

Structured rows reported in this exact model repository at the pinned revision below. They are the publisher talking about itself: never measured evidence, never a score input on their own. They remain quarantined until dataset revision, harness, configuration and identity pass ModelCap’s independent admission review.

Dataset / taskMetricValueRevisionProvenance
cais/hle
hle
—44.1
Unpinned
not published
Reported
Quarantined
datacurve/deep-swe
deep_swe
—72.1
Unpinned
not published
Reported
Quarantined
Delores-Lin/MDPBench
overall
—81.4
Unpinned
not published
Reported
Quarantined
harborframework/terminal-bench
terminalbench_4
—13.1
Unpinned
not published
Reported
Quarantined
harborframework/terminal-bench-2.1
terminalbench_2_1
—81.3
Unpinned
not published
Reported
Quarantined
harborframework/terminal-bench-science
terminalbenchscience_0_1
—5.7
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—93
Unpinned
not published
Reported
Quarantined
likaixin/ScreenSpot-Pro
overall
—83.8
Unpinned
not published
Reported
Quarantined
likaixin/ScreenSpot-Pro
overall
—92
Unpinned
not published
Reported
Quarantined
MathArena/hmmt_feb_2026
MathArena/hmmt_feb_2026
—95.8
Unpinned
not published
Reported
Quarantined
MMMU/MMMU_Pro
mmmu_pro_standard_10_options
—81.6
Unpinned
not published
Reported
Quarantined
ScaleAI/SWE-bench_Pro
SWE_Bench_Pro_hard
—86.3
Unpinned
not published
Reported
Quarantined
ScaleAI/SWE-bench_Pro
SWE_Bench_Pro
—97.7
Unpinned
not published
Reported
Quarantined
SWE-bench/SWE-bench_Multilingual
swe_bench_multilingual_%_resolved
—81.2
Unpinned
not published
Reported
Quarantined
SWE-bench/SWE-bench_Multimodal
swe_bench_multimodal_%_resolved
—43.1
Unpinned
not published
Reported
Quarantined

Other versions (1)

ModelCap Index · Current board

Methodology
56.8
Measured on 5 public benchmarks
Current board
#98 of 371
Score range
50.7–62.9
Rank range
#70–#135
Evidence as of
18 Nov 2026, 14:36 UTC
Benchmarks used
ARC-AGI-2 Verified, ARC-AGI-3 Verified, Arena Coding, Arena Overall, LMArena Agent
Technical details
Public rank basis
Measured (measured)
Evidence label
Measured · 5 public benchmark observations across 5 boards
Index support
48.5% · moderate
Identity binding
Exact catalogue product · thinkingmachines/thinkingmachines/inkling-2-small · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata thinkingmachines/Inkling-2-Small@1df2d97d45210d94f781f42a6521392e5a539e46 (not the evaluation revision)
Index observations used
5
Qualified Index evidence point
56.8
Bayesian Evidence Score · measured evidence input only

BES normalizes admitted public benchmark evidence for measured Index rows. Its legacy score and observables prior do not define the public language rank.

Observed capability
56.8
Evidence status
confirmed · 85% mass
General preference
55.9
Coding
64.2
Agents & tools
54.5
Reasoning
54.5
Evidence breadth
100%
Evidence coverage
49%
Market and catalogue signals · secondary, never the language Index rank
Usage (OpenRouter popularity)
44.9
Liquidity (providers × uptime)
78.6
Open reach (HF downloads)
30.7
Surface (context / tools / modalities)
100.0
Freshness
96.5
  • · OpenRouter popularity #25
  • · 6 live providers
  • · HF 30d downloads 30082
  • · Catalogue: context 1048576, reasoning, tools, modalities Text/Image
  • · First seen 2026-11-09

These adoption and deployment observations remain context only. They do not change this model's ModelCap Index score or rank.

BES evidence selected · measured input only

The bounded 0–100 capability score blends each source’s competitive placement with its published achievement, then combines capability families. Evidence breadth adds a modest uncertainty adjustment; price and popularity are not part of benchmark-led rank.

Deployment readiness

Metadata index
92.6
strong · 94% metadata coverage
Not capability
Artifact reproducibility
85.0
Access & legal clarity
100.0
Deployability
98.7
Evaluation provenance
87.3

Not yet published: the base model it derives from, pinned evaluation dataset versions.

A metadata completeness and deployability index—not a safety certification, quality grade, or production-readiness claim.

Market activity

Venue coverage
Market Gravity
53.8
OpenRouter weekly popularity
#25 of 345

Market Gravity uses OpenRouter’s full-catalogue popularity order. Sparse Vercel top-ten observations appear only when published and do not affect the score.

Arena preference

Status
Ranked
Source
LMArena
Official rank
#152 of 422
Rating
1412.2
95% confidence interval
1400–1424
Votes
3K
Source published at
18 Nov 2026, 00:00 UTC
Category
overall
Identity match confidence
100%

Market Gravity breakdown

Usage55%
44.9

OpenRouter weekly popularity across the full model catalogue.

Liquidity25%
78.6

Independent providers versus the model's open or closed cohort, weighted by uptime.

Open reach15%
30.7

Hugging Face 30-day downloads for open models; neutral for closed models.

Freshness5%
96.5

Time since first public availability, on a six-month half-life.

Specification

Model ID
thinkingmachines/inkling-2-small
Context window
1M tokens
Declared output cap
944K tokens

OpenRouter catalogue or provider-endpoint declaration. Limits can vary by provider and prompt length; a model-wide maximum has not been independently verified. OpenRouter metadata

Inputs
Text, Image
Outputs
Text
Tokenizer
Other
Cached input tokens
$0.03 / 1M
First seen on OpenRouter
9 Nov 2026

Capabilities

  • Supported: Reasoning
  • Supported: Tool use
  • Supported: Structured output
  • Supported: Response format
  • Not supported: Moderated

Weights & access

Open weights. Downloadable weights are published under an OSI or free-culture license in the current repository metadata. This describes weight availability, not whether the full training stack qualifies as Open Source AI.

Access
Open weights
Licence
apache-2.0
Pinned revision
1df2d97d4521
Parameters
594.1B
Architecture
Inkling2ForConditionalGeneration
Model type
inkling2
Hugging Face downloads (30d)
30K
Downloads (all time)
30K
Likes
500
Repository updated
17 Nov 2026, 18:39 UTC

Compare Inkling 2 Small with

Pick any model

Benchmark leaderboards featuring Inkling 2 Small

All leaderboards
Public benchmark boards on which Inkling 2 Small has a published result
LeaderboardScoreSource rankConfigurationPublished
Arena text leaderboardArena (LMArena)14121400–1424#152 of 422Not reported18 Nov 2026
Arena coding leaderboardArena (LMArena)14861462–1509#108 of 417Default18 Nov 2026
LMArena Agent leaderboardLMArena1.5%0.5%–2.5%#26 of 68Default18 Nov 2026
ARC-AGI-2 leaderboardARC Prize Foundation81.7%#57 of 299X-High18 Nov 2026
ARC-AGI-3 leaderboardARC Prize Foundation0.3%#127 of 151X-High18 Nov 2026
Artificial Analysis Intelligence IndexArtificial Analysis30.6#56 of 125Reasoning9 Nov 2026

Alternatives to Inkling 2 Small, by the numbers

Ranked neighbours

Models within 3 places of #98 on the ModelCap Index.

Cheaper at a similar or higher Index

Index score within 3 points of, or above, this model's, with a lower listed blended price (3:1 input:output).

Other ranked Thinking Machines models

The same publisher's other current models with a public Index position.

  • Inkling#62 · Index 65.6 · $4.05/1M out

Inkling 2 Small: common questions

How does Inkling 2 Small rank among AI models?

Inkling 2 Small holds ModelCap Index position #98 of 371 ranked language models as of 19 November 2026, with an Index score of 56.8 (interval 50.7–62.9); evidence: measured. The Index combines public benchmark boards with published uncertainty; the methodology page explains the weighting.

How much does Inkling 2 Small cost per 1M tokens?

Inkling 2 Small is listed at $0.30 per 1M input tokens and $1.00 per 1M output tokens as of 19 November 2026; the lowest output price among 6 providers is $0.50 on PrimeIntellect. Prices come from the OpenRouter catalogue and re-observe on every refresh; at least one provider lists a free tier. For 1,000 requests using 2,000 input and 500 output tokens each, the listed rates imply $1.10 total (2M input + 0.5M output). This excludes caching, batch discounts, prompt-length tiers, tool charges and retries; verify the chosen endpoint's rate and limits before budgeting.

What is the context window of Inkling 2 Small?

Inkling 2 Small has a published context window of 1M tokens (1,048,576). The snapshot lists a 943,718 token catalogue-declared route output cap. OpenRouter catalogue or provider-endpoint declaration. Limits can vary by provider and prompt length; a model-wide maximum has not been independently verified. The providers table lists each endpoint's context limit.

Is Inkling 2 Small open-weight?

Open weights (apache-2.0). Downloadable weights are published under an OSI or free-culture license in the current repository metadata. This describes weight availability, not whether the full training stack qualifies as Open Source AI. The repository is thinkingmachines/Inkling-2-Small on Hugging Face.

Which API providers serve Inkling 2 Small?

6 providers serve Inkling 2 Small through OpenRouter as of 19 November 2026: PrimeIntellect, DeepInfra, Modal, Nebius, Together and Cloudflare. Each provider's price, context limit, quantization and measured uptime are in the providers table above.

Which benchmarks has Inkling 2 Small been evaluated on?

Inkling 2 Small has published results on 6 public boards tracked by ModelCap as of 19 November 2026: Arena text 1412 (#152 of 422), Arena coding 1486 (#108 of 417), LMArena Agent 1.5% (#26 of 68), ARC-AGI-2 81.7% (#57 of 299) and ARC-AGI-3 0.3% (#127 of 151) and more. Each board page ranks every tracked model on that benchmark. The ModelCap Index scores the results its methodology admits, with published uncertainty, and shows reference boards without scoring them.

How does Inkling 2 Small compare with Inkling?

Inkling 2 Small ranks #98 (Index 56.8) and Inkling ranks #62 (Index 65.6) on the ModelCap Index as of 19 November 2026. The head-to-head page lines up their benchmarks, listed prices, context windows and provider counts side by side.

When was Inkling 2 Small released?

Inkling 2 Small first appeared in the catalogue on 9 Nov 2026. It is the current version in its family. Rank and price movements since then are recorded on the site's changes feed.