Inkling · xhigh
Thinking Machines · thinkingmachines/Inkling
Q4 file 587.0 GB (UD-Q4_K_XL) · Weights license: Apache 2.0
Quality
Quality · Overall
LiveBench 2026-06-25 · 0–100 · higher is better · this model outlined, #14 of 22
- Color = publisher
- Memory under review
Source: LiveBench 2026-06-25 · collected 2026-09-26 · open models without a score are not drawn (no score ≠ 0)
No result in LMArena 2026-10-22 Not in LMArena 2026-10-22
LiveBench categories
This model's mean per category · 0–100
Source: LiveBench 2026-06-25 · collected 2026-09-26 · open models without a score are not drawn (no score ≠ 0)
Reported by the publisher
- HLE (text only)
- 29.7%Reasoning
- HLE (with tools)
- 46.0%Reasoning
- AIME 2026
- 97.1%Reasoning
- GPQA Diamond
- 87.2%Reasoning
- SWEBench Verified
- 77.6%Agentic (coding)
- SWEBench Pro (Public)
- 54.3%Agentic (coding)
- Terminal Bench 2.1 (Best Harness)
- 63.8Agentic (coding)
- GDPVal-AA v2
- 1,233Agentic (coding)
Every reported result
- MCP Atlas
- 74.1%Agentic (general)
- Tau 3 Banking
- 23.7%Agentic (general)
- BrowseComp (w/ Ctx)
- 77.1%Factuality
- SimpleQA Verified
- 43.9%Factuality
- AA Omniscience
- 1.0%Factuality
- IFBench
- 79.8%Chat
- Global-MMLU-Lite
- 88.7%Chat
- MMMU Pro (Standard 10)
- 73.5%Vision
- Charxiv RQ
- 78.1%Vision
- Charxiv RQ (with python)
- 82.0%Vision
- Audio MC
- 56.6%Audio
- MMAU
- 77.2%Audio
- VoiceBench
- 91.4%Audio
- FORTRESS (Adversarial)
- 78.0%Safety
- FORTRESS (Benign)
- 95.9%Safety
- StrongREJECT
- 98.6%Safety
Values as printed on the card, with the publisher's own settings (card section: 5. Evaluations); column “Inkling”. Not comparable across publishers and not used in any ranking. Model card
Run it
Run it
File UD-Q4_K_XL · 587 GB · 14 parts · unsloth/inkling-GGUF · What is Q4?
llama-server -hf unsloth/inkling-GGUF:UD-Q4_K_XL- Install llama.cpp
- Run the command above
- Ask at localhost:8080:
Explain what a context window is, in two sentences.
Quality × memory
Quality × memory
LiveBench 2026-06-25 overall vs. Q4 planning memory (log scale) · 18 open models
- This model
- Color = publisher
Memory = the model's own figure: parameters × 0.625 bytes + 20% working room, before any machine's reserve; with a chosen machine, the total there with its reserve (25% of memory, 2 to 8 GB; 2 GB per graphics card). An estimate, not measured VRAM. LiveBench 2026-06-25
Runs on
More memory than any profile in the catalog
Even the largest profile (8 × 80 GB, aggregated) is below this planning estimate. Multi-node setups are outside this catalog.
Inkling · xhigh is a text model from Thinking Machines with 975B parameters, 41B of them active per token (mixture of experts). It was released on July 15, 2026 and reads up to 1,000,000 tokens of context. Its Q4 weights plus working room come to about 732 GB, more than the fast memory of every reference machine. We collected 8 downloadable GGUF and MLX files from 2 publisher sources; the most downloaded GGUF source is unsloth's. It scores 71.9 overall in LiveBench 2026-06-25.
On the reference machines
| Where Q4 sits | ||||
|---|---|---|---|---|
| Mac mini M2 · 8 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~734 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| Mac mini M4 · 16 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~736 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| Mac mini M4 · 24 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~738 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| Mac mini M4 · 32 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~740 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| Mac mini M5 Pro · 64 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~740 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| Mac Studio M5 Max · 128 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~740 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| Mac Studio · M5 Ultra · 36-core CPU / 80-core GPU · 512 GB unifiedLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~740 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| RTX 4060 · 8 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~734 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| RTX 3060 · 12 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~734 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| RTX 5080 · 16 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~734 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| RTX 4090 · 24 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~734 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| RTX 5090 · 32 GB VRAMLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~734 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
| DGX H100 · 8 × 80 GBLarger than its memoryNo file on the site's quant ladder fits the fast memory at 8,192 tokensNo file in the previous column | ~748 GB | Larger than its memory | No file on the site's quant ladder fits the fast memory at 8,192 tokens | No file in the previous column |
The file column uses the site's quant ladder (Q8_0 to Q3_K_M, MLX 8 to 4-bit on a Mac): the largest collected file whose size × 1.2 plus the context cache and the machine's reserve fits the fast memory. Planning estimates, not measured runs.
Compare Inkling · xhigh using its publisher card and the LiveBench 2026-06-25 snapshot. This is quality evidence, not local speed.
The Q4 weights plus working room come to about 732 GB from 975B stated parameters, before the machine's reserve (Runs on adds it per machine). It is not a measured minimum; context, batch size, runtime and quantization change actual memory use.
Versions to download
Files you can download for this checkpoint. The section opens on the most downloaded collected GGUF source, unsloth, at UD-Q4_K_XL. On a Mac it opens on the most downloaded MLX source. Sizes are the Hub's published bytes. What is Q4?
unsloth
unsloth/inkling-GGUF · UD-Q4_K_XL/inkling-UD-Q4_K_XL-00001-of-00014.gguf · 587.0 GB
Run this file
llama-server -hf unsloth/inkling-GGUF:UD-Q4_K_XLlms get https://huggingface.co/unsloth/inkling-GGUF- Ollama: Ollama's documented hf.co form covers single-file GGUFs; this file has 14 parts
- mlx-lm: mlx-lm runs MLX repositories; pick the MLX version on a Mac
Quality scores on this page belong to the original checkpoint; these files are not scored or ranked. File sizes read 2026-10-23 · download counts read 2026-10-17 · commands: Hugging Face Hub docs: Use Ollama with any GGUF model · llama.cpp server README (-hf <user>/<model>[:quant]) · LM Studio docs: lms get with a Hugging Face URL · mlx-lm README (mlx_lm.generate --model)
Fine-tunes, adapters and other versions
- Quantized versions
- 48
- Fine-tunes
- 15
- Adapters · merges
- 12 · 0
Browse them on the Hugging Face model tree · read 2026-10-23
Uncensored and abliterated versions
Some community uploads remove the model's safety refusals. Quality may drop, the original publisher does not support them, claims about them are the uploader's, and the original license still applies. This site never recommends or ranks them.
1 derivative repositories have “uncensored” or “abliterated” in their name (read 2026-10-23).
Measured speed
No published measurement for this checkpoint
Tokens per second are shown only when a source publishes them with hardware, build and setup. They are never estimated.
How to run
The memory needed above is the Q4 load before a machine's reserve. Formula: parameters × 0.625 bytes, plus 20% working allowance; each machine then adds its system reserve: 25% of unified or system memory, 2 to 8 GB, or 2 GB per GPU on dedicated cards. This is neither measured VRAM nor checkpoint download size.
Q4 planning estimate only. Check checkpoint format, runtime support, KV cache, context, batch size and offload before choosing hardware.
Evidence level: Generic estimate only; no execution reproduced by this site.
- Publisher sampling defaults
- — generation_config.json states no sampling values for this checkpoint
- Chat template markers
- tool calls · thinkingmarkers found in the template, not a test of the feature · read 2026-10-23source
Before downloading
- Check the exact checkpoint and license linked here.
- Confirm a compatible runtime and precision or quantization for your platform.
- Allow for context, batch size and any extra media components beyond the weights.
Technical details
What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.
- Checkpoint / version
- thinkingmachines/Inkling
- Task and modality
- language and reasoning · text + image + audio
- Count and basis
- 975B · manufacturer specification · 41B active
- Context
- 1,000,000 tokens published for this model
- Architecture
- MoE
- Released
- Inkling Benchmarks, Cost and Capabilities (Vals AI) · read 2026-10-09secondary source
- Runtime and precision
- No local execution verified.
- Languages
- en Publisher's statement · read 2026-10-18source
- Commercial use
- Allowed Apache License 2.0 text (SPDX) · read 2026-10-18license text
Every stored field
- Revision checked
- 828496eeae4c243ff1a22f7f28ff83694f2f7bc9
- Architecture
- InklingForConditionalGeneration
- Library
- transformers
- Hub task tag
- image-text-to-text
- Formats
- safetensors
- Repository created
- 2026-07-14creation is not the model's release
- Repository updated
- 2026-07-23a commit is not a release date
- Hub downloads (30 days) · likes
- 1,587,519 · 1,827this repository only · read 2026-10-25
- Read from the Hub page
- task, transformers class, task tags, file formatsHub overview and file listing · read 2026-10-18
- Catalog state
- Hub config or weights found; no inventory benchmark link
Manufacturer guidance: at least 2 TB VRAM for BF16 or 600 GB for NVFP4. Those configurations differ from this tool’s generic Q4 planning estimate.
Sources
Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.
- Publisher checkpoint and model card
- Weight license and use terms
- Quality benchmark source · LiveBench 2026-06-25 overall
Catalog reviewed: 2026-09-30. Benchmark release: 2026-06-25; snapshot collected: 2026-09-26.