Skip to content
aiqbench

Thinky · model passport

Thinking Machines: Inkling Small

inkling-small
  • Open weights
  • Apache 2.0 licence
  • 266.0B params
  • 524,288 context
  • Released 2026-07-27(HF repo created)
12/27axes covered
Overall · LMArena
1,404#143 of 396
Best percentile
77.4GPQA Diamond
Cohort standing
51.2median 60.1

What the data says

Benchmarked on 12 of 27 tracked axes. Every figure below is the model's own published result on that board.

Weakest axes

  • 3.6 pct — Agent Arena (-0.147)
  • 13.4 pct — SimpleQA Verified (19.1%)

Uneven: 73.8 percentile points between its best and worst axis — 77.4 pct on GPQA Diamond versus 3.6 pct on Agent Arena.

  • GPQA Diamond77.4 pct88.5%Verify →
  • Coding72.4 pct1,473Verify →
  • AIME (math)71.5 pct90.0%Verify →
  • Hard prompts64.9 pct1,426Verify →
  • Overall64.1 pct1,404Verify →
  • Instruction following63.6 pct1,394Verify →
  • ARC-AGI-258.4 pct40.1%Verify →
  • Vision Arena50 pct1,208Verify →
  • WebDev Arena42.4 pct1,409Verify →
  • Factuality Arena35.5 pct1,424Verify →
  • SimpleQA Verified13.4 pct19.1%Verify →
  • Agent Arena3.6 pct-0.147Verify →
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • FrontierMathNot benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • IFBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

Repositorythinkingmachines/Inkling-Small
Parameters266.0B
ArchitectureInklingForConditionalGeneration
LicenceApache 2.0

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "thinkingmachines/Inkling-Small"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Use it hosted

API model IDthinkingmachines/inkling-small
Context length524,288 tokens
Max output262,144 tokens
Acceptstext, image, audio
Input price$0.45 per 1M tokens
Cached input price$0.10 per 1M tokens
Output price$1.20 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogit biasMax tokensMin-pPresence penaltyReasoningReasoning effortRepetition penaltySeedStop sequencesTemperatureTool choiceTool callingTop-kTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"thinkingmachines/inkling-small","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • Inkling Small stands at 51.2 (raw mean 51.4) — measured on 12 of 27 axes.

    standingtemplateVerify →

  • Inkling Small stands at 51.2 versus a cohort median of 60.1 across 194 models released within 180 days — below the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →