Inkling
Intelligence benchmarks
Artificial Analysis indexes - compared with the best open and proprietary models
Intelligence
25.0
AA Index
Coding
52.1
AA Index
Agentic
22.5
AA Index
Intelligence Index - Inkling vs. the field
Best open-weight models (you can run locally) and leading proprietary models for context.
Coding Index comparison
Agentic Index comparison
Benchmark data from Artificial Analysis · updated 2026-10-15.
Hosted API cost & speed
What it costs to rent this model instead of running it locally
| Metric | Value |
|---|---|
| Input (per 1M tokens) | $1.00 |
| Output (per 1M tokens) | $4.05 |
| Median speed (hosted API) | 180 tok/s |
Median across hosted providers, measured by Artificial Analysis - updated 2026-10-15. Local llama.cpp speed depends on your hardware — see the hardware pages for estimates.
Arena rating (LMArena)
Human preference wins from head-to-head chat battles
| Metric | Value |
|---|---|
| Overall arena rating | 1441.2 |
| Rank (overall) | #73 |
| Battle votes | 39,763 |
Overall rating published by LMArena on 2026-10-13 - a pairwise human-preference Elo across live chat battles; higher is better. Fetched 2026-10-15. Open-weights models compete head-to-head with proprietary ones here — no benchmark prompts, just what people pick.
Standard benchmarks
Performance across standard evaluations
| Benchmark | Score |
|---|---|
| GPQA Diamond | 87.2 |
| HLE | 31.9 |
| SciCode | 47 |
Source: Artificial Analysis component evals - updated 2026-10-15.
Will it run on your hardware?
Pick your GPU memory - see which quantizations fit, and the cheapest card for the rest
Need an exact figure for your context length? Use the VRAM calculator.
Run it locally
Copy-paste - running in under a minute
vllm serve thinkingmachines/Inkling-NVFP4New to this? Start with Ollama · serve to many users with vLLM.
Deep dive
Notes, sources, and the full write-up
Inkling is a 552.8B-parameter apache-2.0 model from Thinking Machines. It scores 25 on the Artificial Analysis Intelligence Index (coding 52.1). At Q4_K_M it needs roughly 321 GB of VRAM, placing it in the 48 GB+ / multi-GPU hardware tier.
Benchmarks
Artificial Analysis Intelligence Index - Inkling vs. leading closed models:
| Model | Intelligence | Coding | GPQA |
|---|---|---|---|
| Inkling | 40.7 | 52.1 | 87.2 |
| Claude Fable 5 (with fallback) | 59.9 | 76.5 | 92.6 |
| GPT-5.6 Sol (max) | 58.9 | 77.4 | 94.1 |
| Kimi K3 | 57.1 | 76.2 | 93.5 |
| Claude Opus 4.8 (max) | 55.7 | 74.3 | 92 |
| GPT-5.6 Terra (max) | 55 | 76.7 | 92.5 |
Source: Artificial Analysis (2026-07-18).
Popularity
Inkling has 38,374 downloads in the last month on HuggingFace and 61 likes.
Frequently asked
Quick answers to common questions
How much VRAM does Inkling need?
Inkling with 552.8B parameters needs approximately 321 GB at Q4_K_M quantization. Use our VRAM calculator for an exact estimate.
Is Inkling better than other thinkingmachines models?
Inkling has 552.8B parameters with 8,192 context - a strong choice for general use.
What license is Inkling under?
Inkling is released under the apache-2.0 license, making it suitable for most commercial and personal projects.
What hardware runs Inkling well?
With 552.8B parameters, Inkling requires adequate VRAM. High-end GPUs like the RTX 4090 (24GB), RTX 5090 (32GB), or Mac Studio with unified memory are good options. Check our hardware directory for specific recommendations.
What is the best quantization for Inkling?
Q4_K_M is the recommended sweet spot - ~98% of FP16 quality at ~27% of the size. Q5_K_M (~392 GB) is an option if you have spare VRAM. Use our VRAM calculator to compare.
What models compete with Inkling?
Inkling competes with other 276B–829B. Browse our model directory for comparisons, benchmarks, and community reviews to find the best fit.
Nearby options
Similar models and compatible hardware by spec