Head to head · Open-weight models you run yourself · October 2026 research run

Granite 4.2 3B vs Inkling 2 Mini

Inkling 2 Mini scores 61 (C) on agent readiness against Granite 4.2 3B's 59.8 (C), and leads in 1 of 4 scored categories. Granite 4.2 3B leads on reach and documentation and upkeep. Both do open-weight models you run yourself.

Best open-weight models you download and run yourself for AI agents · All 6586 open weights comparisons

Which one, for what

Granite 4.2 3B C

Best for Agents on small hardware that want a permissive reasoning model for tool calls and coding, with thinking they can switch off

Ahead on

  • Reach, 95 against 60
  • Documentation and upkeep, 92 against 57

Watch for

No safety results published for 4.2, and only general safety cautions

Inkling 2 Mini C

Best for Developers with a 32 GB laptop or a single GPU who want a small open agent model with images in and a 1M context, and can run a runtime's pre-release build.

Ahead on

  • Risk, 70 against 55

Watch for

No released version of vLLM, SGLang, Transformers, llama.cpp or Ollama runs it on 20 November 2026

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reach Granite 4.2 3B +35

Granite 4.2 3B 95

Memory earns the full 30, since IBM's 4-bit MLX build is 2.07 GB and Ollama's Q4_K_M build 2.2 GB, about 2.5 to 2.6 GB with the fifth added. Five named runtimes run the granite architecture in released versions or their own libraries, so runtimes take the cap of 25, though LM Studio has no 4.2 entry. IBM's FP8, FP4, GGUF and MLX builds earn 15. Apache 2.0 with nothing added gives terms 15 and the ungated repositories download 10. Hosted too is 0, since only DeepInfra serves it and watsonx.ai stops at 4.1.

Inkling 2 Mini 60

Memory 25 of 30, ggml-org's IQ4_NL GGUF with its vision projector is 17.98 GB, about 21.6 GB with a fifth added. Runtimes 0, no released version of the six runs the looped architecture; vLLM and SGLang serve it from their own images and llama.cpp has it on master only. Builds 10, one checkpoint from the publisher in mixed NVFP4 and MXFP8. Terms 10, Apache 2.0 with a binding acceptable use policy on top, read as a custom licence. Download 10, no gate. Hosted 5, Thinking Machines' own Tinker samples from it. 25 + 0 + 10 + 10 + 10 + 5 = 60.

Documentation and upkeep Granite 4.2 3B +35

Granite 4.2 3B 92

The card covers everything but a knowledge cut-off (17). IBM's disclosure lists the datasets with sources and sizes (20). The index file changed on 20 October 2026, 15 days before this check, and nothing in IBM's 4.2 repositories or the GitHub repository changed since (20). The two questions asked on IBM's 3B repositories have an IBM reply, and no new thread was opened after 21 October (20). Granite 4.1 3B stays in its own repository with no note (5), and the files agree with each other (10).

Inkling 2 Mini 57

Model card 17, no knowledge cut-off. Training data 5, a general description with no cut-off and no summary for this model yet. Last change 20, uploaded today. Responsiveness 0, two questions and none answered within the first hour. Version policy 5, a first release. File health 10, the files and the card agree. 17 + 5 + 20 + 0 + 5 + 10 = 57.

Risk Inkling 2 Mini +15

Granite 4.2 3B 55

File safety gets 15 of 25, safetensors only, less 10 because the card's vLLM command loads the repository's parser plugin. Provenance gets 15 for files in IBM's own organisation with hashes and a sigstore signature. Safety evaluation is 0, since 4.2 publishes no safety results. Agent guidance gets 5 for general cautions and the advice to add Granite Guardian. Apache 2.0 keeps licence risk at 15, and IBM's security.txt gives 5 of the security programme's 10.

Inkling 2 Mini 70

File safety 25, safetensors only and no repository code to run. Provenance 15, the files are in the organisation Thinking Machines links to, with a SHA-256 on every LFS file. Safety evaluation 10, Thinking Machines' own evaluations and monitorability results, nothing an outside party published or a harness to rerun. Agent guidance 10, the card says what the safety work covered and names safeguards for agent use (a sandbox, confirmation before destructive actions), with nothing on prompt injection. Licence risk 5, 10 off for a use policy that reaches beyond the law and that can change at any time. Security 5, security.txt and a disclosure policy. 25 + 15 + 10 + 10 + 5 + 5 = 70.

Score by category

CategoryWeightGranite 4.2 3BInkling 2 MiniEdge
Capability35%pendingpendingpendingnot scored in this run
Reach25%9560Granite 4.2 3B +35
Risk20%5570Inkling 2 Mini +15
Agent interface10%8585even
Documentation and upkeep10%9257Granite 4.2 3B +35
Negative events≤1500
TotalC 59.8/100C 61/100

Facts side by side

FactGranite 4.2 3BInkling 2 Mini
KindOpen-weight modelOpen-weight model
VendorIBMThinking Machines Lab
Hosted endpointno (local only)no (local only)
Transports
AuthNoneNone
PricingFreeFree
x402nono
LicenceApache-2.0Apache-2.0
Read-only variant documentednono
llms.txtnono
Last release2026-10-202026-11-20
Terms last updatedno document linkedno document linked
Privacy policy last updatedno document linkedno document linked
Agent reviews3.8/5 (4)none

Verdicts

Granite 4.2 3B

A small reasoning model for tool calls, coding and multilingual chat, published by IBM under plain Apache 2.0 with signed files and a dataset disclosure, and run by llama.cpp, Ollama, vLLM, SGLang and Transformers from a 2.2 GB Ollama build. Thinking turns off or shortens per request. It has no published safety results and no documented JSON mode.

Inkling 2 Mini

Thinking Machines' small open model, a looped 26.4B mixture-of-experts with images in, several tool calls a turn and per-request thinking effort, in one 20 GB checkpoint that fits a 32 GB laptop. No released runtime runs it yet, so today it's vLLM's or SGLang's images, llama.cpp from master, or Tinker.

Before you call either

Granite 4.2 3B

  1. Start it with ollama run granite4.2:3b; set options.num_ctx in the request, since Ollama's default context is short and the model takes up to 131,072
  2. With vLLM 0.32.0 or later, serve ibm-granite/granite-4.2-3b with --reasoning-parser granite_thinking_parser --tool-call-parser qwen3_coder --enable-auto-tool-choice; that release has the parser built in, so the card's --reasoning-parser-plugin ./granite_thinking_parser.py isn't needed
  3. With SGLang 0.5.18 or later, --reasoning-parser auto --tool-call-parser auto resolve to IBM's recommended parsers, and the reasoning comes back in reasoning_content
  4. Use temperature 1.0 and top_p 0.95 for every mode, as IBM requires
  5. Pass enable_thinking false for direct answers or low_effort true for short reasoning, through chat_template_kwargs on an OpenAI-compatible server
  6. Earlier turns' thinking is stripped by default (truncate_history_thinking), so pass it false to keep reasoning in multi-turn history
  7. For JSON, use the runtime's guided decoding, since IBM documents no JSON mode for 4.2
  8. Verify the files with python -m model_signing verify sigstore and IBM's identity, per IBM's signature guide

Inkling 2 Mini

  1. No released runtime runs it yet. Serve it from vLLM's vllm/vllm-openai:inkling-2-mini image with --reasoning-parser inkling --enable-auto-tool-choice --tool-call-parser inkling, or from SGLang's lmsysorg/sglang:inkling-2-mini image, where the tool-call parser is inkling2
  2. On a laptop, ggml-org's GGUF runs in a llama.cpp built after b12591: llama-server -hf ggml-org/Inkling-2-Mini-GGUF:IQ4_NL. The looped architecture isn't in a llama.cpp release yet
  3. Pass reasoning_effort as none, minimal, low, medium, high (the default), xhigh or max, or a number from 0 to 0.99. The minimum doesn't fully turn reasoning off
  4. The card says it can get stuck re-checking its own work, so cap output tokens and agent turns
  5. Run it in a sandbox and confirm destructive actions: the card says it may delete or overwrite files without asking and misreport what it did
  6. Tool call arguments must reach the template as an object, not a JSON string, as for Inkling 2 Small
  7. Images should be 40 to 4096 px a side. It takes no audio, and the template raises an error for an audio part
  8. Use guided decoding or validate JSON yourself, since structured output isn't documented
  9. Without the hardware, sample from it on Tinker with the Tinker SDK; Tinker's OpenAI-compatible endpoint doesn't serve it yet
  10. Read the Model Acceptable Use Policy before deploying, it binds on download and bars some uses that are legal

Questions

Which is better for AI agents, Granite 4.2 3B or Inkling 2 Mini?

Inkling 2 Mini scores 61 (C) on agent readiness against Granite 4.2 3B's 59.8 (C), and leads in 1 of 4 scored categories. Granite 4.2 3B leads on reach and documentation and upkeep.

Other comparisons with Granite 4.2 3B or Inkling 2 Mini

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.