# Inkling 2 Mini: a looped transformer you can run on a laptop

> Thinking Machines' new open-weights model loops its middle layers three times per token. It leads open-weights models under 100B on agentic coding, runs in about 20GB, and is live on AnyRouter through Tinker and OpenRouter BYOK.

- Date: November 21, 2026
- Category: product
- HTML: https://anyrouter.dev/blog/inkling-2-mini

## A small model that thinks more per token

Inkling 2 Mini is a 26.4B-parameter mixture-of-experts model with 3.4B active parameters. Its unusual trick is looping: the middle layers of the network run three times for every token, so the model does more serial computation without storing more parameters. Thinking Machines reports that it leads every open-weights model under 100B total parameters on agentic coding, including 48.2% on DeepSWE v1.1 and 46.5% on the Vals Index v2.2.

It runs natively in NVFP4 and its whole footprint is about 20GB. That is small enough for a consumer laptop: the lab ran it offline on an M5 Pro MacBook Pro and built a playable flight simulator in a few minutes.

**DeepSWE v1.1 — tasks resolved**

- Inkling 2 Mini: 48.2% %
- Qwen3.8-27B: 42.2% %
- Nemotron 3.5 Super VL: 24.7% %
- Muse Glimmer: 4.4% %
- Gemma 4 31B: 0.3% %

_Figure 1: DeepSWE v1.1 scores from the Thinking Machines comparison table._

## The benchmark table, made interactive

The launch post compares Inkling 2 Mini with Gemma 4 31B, Muse Glimmer, Qwen3.8-27B and Nemotron 3.5 Super VL across coding, agentic search, vision, reasoning and safety. We copied every published number into a compare collection. Switch between the table and the Pareto and saturation charts, sort by model, and share a direct link to any view.

**Inkling 2 Mini vs open-weight peers**

| Vendor benchmarks | Inkling 2 Mini | Gemma 4 31B | Muse Glimmer 30B | Qwen3.8-27B | Nemotron 3.5 Super VL 120B A12B |
| --- | --- | --- | --- | --- | --- |
| Model info |  |
| Inkling 2 Mini has 3.4B unique active parameters. The vendor's 'activated' column below counts weight activations per token (5.4B), which includes the three loop passes. |  |
| Params (B) (activated per token (incl. loops) / total) | 5.4/26.4 | **30.7/30.7** | 30.0/30.0 | 27.0/27.0 | 12.0/124.5 |
| Third party index |  |
| AA Index (v5.0) | **20.1** | 4.4 | 6.3 | **20.1** | 13.6 |
| Vals Index (v2.2) | **46.5%** | – | – | – | 29.8% |
| Coding and autoresearch |  |
| FrontierCode v2.0 (stirrup v0.2.1) | **15.1%** | 0.0% | 1.5% | 14.9% | 6.1% |
| DeepSWE v1.1 (mini-swe-agent v2.4.6) | **48.2%** | 0.3% | 4.4% | 42.2% | 24.7% |
| Terminal Bench v4.0 (mini-swe-agent v2.4.6) | **5.6%** | 0.0% | 0.5% | **5.6%** | 0.0% |
| Terminal Bench v2.1 (Inkling 2 Mini in pi v0.83.0) | 71.6% | 43.4% | 51.7% | **73.0%** | 60.7% |
| SWEBench Pro v2 (Hard) (mini-swe-agent v2.4.6) | **77.9%** | 14.2% | 44.3% | 73.8% | 44.1% |
| SWEBench Multilingual (Inkling 2 Mini with bash only) | **73.7%** | 43.0% | 66.9% | 70.0% | 67.7% |
| SWEBench Verified (bash only) | **79.2%** | 46.1% | 76.0% | 77.8% | 71.1% |
| Vibe Code Bench v1.1 (openhands (vals)) | **77.9%** | – | – | 64.9% | 3.5% |
| PMPP-Hard (AUC, 0-10h) (pi v0.83.0) | **455.8** | – | – | 262.5 | – |
| Agentic search, tool use and knowledge work |  |
| BrowseComp (Inkling 2 Mini with compaction) | **79.0%** | 37.0% | 63.9% | 61.6% | 78.9% |
| DeepSearchQA (Inkling 2 Mini with compaction) | **85.3%** | 61.7% | 74.6% | 79.7% | 77.7% |
| MCP Atlas (Public) | **82.7%** | 54.2% | 75.5% | 77.3% | 67.7% |
| AA-Briefcase v1.1 | 1263.0 | 369.0 | 478.0 | **1391.0** | 999.0 |
| GDPVal-AA v2.1 | 1319.0 | 627.0 | 792.0 | **1427.0** | 1162.0 |
| JobBench | **49.6%** | 8.2% | 16.1% | 33.4% | 36.5% |
| AutomationBench-AA v1.0.6 | 41.8% | 4.9% | 6.8% | **48.2%** | 14.8% |
| EnterpriseOps-Gym | 42.0% | 28.4% | 34.7% | **44.2%** | 28.1% |
| Vals Finance Index | **50.1%** | – | – | – | 41.0% |
| AA-AnalystAgent | **46.3%** | – | 20.0% | – | – |
| Diligence Bench (best harness) | **48.6%** | – | 40.9% | 46.4% | 45.0% |
| Vision interpretation and creation |  |
| GDP.pdf | **16.8%** | 6.0% | 11.8% | 16.4% | 8.0% |
| PerceptionBench | **57.5%** | 40.7% | 42.8% | 52.4% | 36.8% |
| MMMU Pro (standard 10) | 74.9% | 73.4% | 74.3% | **76.3%** | 72.3% |
| OmniDocBench v1.6 | **92.8%** | 82.5% | 90.4% | 91.5% | 77.9% |
| ScreenSpot Pro (original / with Python) | **76.5/87.7%** | 5.9%/– | 75.4%/– | 57.6/78.9% | 43.5%/– |
| Chartography (original / with Python) | **22.1/72.5%** | 12.0/11.8% | 17.4/25.1% | 12.2/52.1% | 9.7/29.3% |
| ZeroBench (main) (original / with Python) | **17.0/34.0%** | 9.0/12.0% | 9.0/17.0% | 12.0%/– | 11.0/12.0% |
| MuirBench | 77.5% | 72.3% | **77.6%** | 65.9% | 72.7% |
| BVB | 68.4% | 47.0% | 50.5% | **70.7%** | 52.1% |
| Reasoning, knowledge and epistemics |  |
| HLE-Diamond | 15.2% | 7.2% | 9.1% | **15.3%** | 8.8% |
| CritPt v2 (full / hard) | 31.4/17.0% | –/4.0% | –/7.0% | 10.6/4.0% | **36.9/13.5%** |
| Apex 2025 | **34.4%** | 5.2% | 4.2% | 18.8% | 22.9% |
| ARC-AGI (v1 / v2) | 84.5/49.3% | – | – | **87.5/42.4%** | 22.7/1.5% |
| LongBench v2 | **66.9%** | 52.9% | 33.1% | 52.3% | 62.1% |
| GraphWalks BFS (256K) | **69.7%** | – | – | 50.4% | 63.3% |
| MRCR v2 (8-needle 256K-512K) | **46.4%** | – | – | 38.0% | 5.7% |
| AA-Omniscience v1.1 (score) | -8.6 | -48.5 | -32.2 | -6.4 | **-2.9** |
| AA-Omniscience v1.1 (non-hallucination rate) | 63.4% | 14.5% | 16.4% | **71.4%** | 69.5% |
| ComplexConstraints v1.1 (openhands v1.45.0) | 33.4% | 31.6% | **35.6%** | 30.1% | 30.1% |
| Safety and alignment |  |
| FORTRESS (adversarial public) | 89.5% | 67.8% | **90.3%** | 78.8% | 79.6% |
| FORTRESS (benign public) | 96.4% | 97.6% | 89.1% | **97.8%** | 92.7% |
| StrongREJECT (none) | 97.8% | 98.1% | 99.0% | 99.0% | **99.4%** |
| Misbehavior (internal) (lower is better) | **9.8%** | 27.6% | 46.4% | 57.7% | 52.8% |
| Covert misbehavior (internal) (lower is better) | **8.3%** | 24.3% | 45.6% | 33.2% | 43.4% |
| Destructive actions (internal) (lower is better) | **10.2%** | 75.8% | 68.8% | 35.2% | 65.6% |

| Independent evals | Inkling 2 Mini | Gemma 4 31B | Muse Glimmer 30B | Qwen3.8-27B | Nemotron 3.5 Super VL 120B A12B |
| --- | --- | --- | --- | --- | --- |
| GPQA Diamond (Artificial Analysis) | – | 76.3% | – | – | – |
| ARC-AGI-2 (Epoch AI) | – | – | – | 42.4% | – |
| CursorBench | – | – | – | – | – |
| SciCode (Artificial Analysis, Epoch AI) | – | 41.1% | – | **46.6%** | – |
| SimpleQA (verified) | – | – | – | – | – |

_Sources: [Introducing Inkling 2 Mini (Thinking Machines Lab)](https://thinkingmachines.ai/news/inkling-2-mini/) (as of 2026-11-21); [Epoch AI](https://epoch.ai/benchmarks) (CC-BY 4.0). Interactive version: https://anyrouter.dev/compare/inkling-2-mini_

## What looping actually changes

The model has 24 layers in four blocks of six. The first and last blocks run once; the two middle blocks run three times. The lab compared loop position on its scaling ladder and found that looping the middle layers bought more compute-equivalent gain than looping the last layers.

| Loop position | Compute-equivalent gain | Layer span |
| --- | --- | --- |
| Middle layers | 1.51–1.91× | Blocks 2 and 3, layers 6–18 |
| Last layers | 1.24–1.36× | Last 12 layers |

_Figure 2: Compute-equivalent gain by loop position at training step 2,000._

The attention-residual design lets each pass read the outputs of earlier blocks, and the loop's message is gated so the model learns when to use it. Roughly 20 of the 24 layers are deployed in NVFP4 and the final four in MXFP8, the same precision used during training, so there is no quantization conversion loss.

## Use it on AnyRouter

Inkling 2 Mini is available as `thinkingmachines/inkling-2-mini`. Both routes are BYOK: add a Tinker key or an OpenRouter key under Dashboard → BYOK and your provider bills you directly. AnyRouter adds no charge for a BYOK route. Tinker lists the 256K train / 1M sample wire at $0.15 per 1M input tokens and $0.50 per 1M output tokens.

```ts
import { createAnyRouter } from "@anyr/ai-sdk-provider"
import { generateText } from "ai"

const anyrouter = createAnyRouter()
const { text } = await generateText({
  model: anyrouter("thinkingmachines/inkling-2-mini"),
  prompt: "Explain the difference between looping and chain of thought.",
})
console.log(text)
```

_AI SDK._

```bash
anyr claude --model "thinkingmachines/inkling-2-mini"
```

_Coding agent._

## Where it still falls short

The lab is candid about the limits. It is not the strongest small model for multi-turn chat or instructions with many constraints, scoring 50.0% on MultiChallenge and 33.4% on ComplexConstraints v1.1. Looping adds reasoning depth but not stored facts, so factuality stays in line with other same-size models, and it trails larger models on reasoning-heavy benchmarks such as CritPt v2 and HLE-Diamond. For knowledge work it leans on search and tools instead.
