- Size
- 266B parameterscounted from the safetensors header on the Hub
- Licence
- custom, read it, Apache License 2.0Our reading, check the licence: A custom licence we have not read for this model: other makers use the same name for different texts. Read the model's own licence file before you use it.Licence text
- On the Hub since
- 27 Jul 2026the date the repo was created
- Context length
- 512k tokensas an API provider reports it
- Kind
- Base model (no base model declared on the Hub)InklingForConditionalGeneration
- On Hugging Face
- thinkingmachines/Inkling-Small963k downloads in the last 30 days, 458 likes
Scores
- effort xhigh17.1%±5.914 Aug 2026, log
Epoch's name: inkling-small. Best closed model on this set: claude-fable-5-2_max, 100.0% (Anthropic). All 83 results
- effort xhigh46.3%±3.014 Aug 2026, log
Epoch's name: inkling-small. Best closed model on this set: claude-fable-5-2_max, 95.1% (Anthropic). All 127 results
- effort xhigh88.5%±1.914 Aug 2026, log
Epoch's name: inkling-small. Best closed model on this set: gpt-6.2-astra_max, 96.8% (OpenAI). All 341 results
- effort xhigh6.0%±2.415 Aug 2026, Epoch AI
Epoch's name: inkling-small. Best closed model on this set: gpt-6.2-astra_max, 95.0% (OpenAI). All 148 results
- effort xhigh90.0%±4.014 Aug 2026, log
Epoch's name: inkling-small. Best closed model on this set: claude-fable-5_high, 100.0% (Anthropic). All 318 results
- effort xhigh19.1%±1.227 Aug 2026, log
Epoch's name: inkling-small. Best closed model on this set: claude-fable-5-2_max, 94.0% (Anthropic). All 105 results
- effort xhigh84.0%arcprize.org
- effort high78.0%arcprize.org
- effort medium67.0%arcprize.org
- effort low52.5%arcprize.org
- effort minimal48.5%arcprize.org
- effort none5.5%arcprize.org
Epoch's name: inkling-small. Best closed model on this set: claude-fable-5_max, 98.5% (Anthropic). All 318 results
- effort xhigh40.1%via Epoch AI
- effort high33.1%via Epoch AI
- effort medium13.6%via Epoch AI
- effort low4.9%via Epoch AI
- effort minimal4.0%via Epoch AI
- effort none0.0%via Epoch AI
Epoch's name: inkling-small. Best closed model on this set: gpt-6-astra_max, 95.0% (OpenAI). All 299 results
- default setting8.3%via Epoch AI
Epoch's name: inkling-small. Best closed model on this set: gemini-4-carbon_high, 36.0% (Google DeepMind). All 271 results
- Reasoning effort: 0.996.0%±2.4vals.ai
Epoch's name: inkling-small. Best closed model on this set: claude-fable-5-1_max, 100.0% (Anthropic). All 85 results
Claimed by the maker Claimed
The model card has no claimed scores in a machine-readable form (the card's model-index), so none are shown. The maker may give scores in the card text: read the card.
Other views, which we link to and do not copy: look this model up on Artificial Analysis and the Arena leaderboard.
Independent scores from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Each row links its own source. What the badges mean.
Run it yourself
Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 1.3 GB, worked out from the model's config (thinkingmachines/Inkling-Small), at 16-bit precision. Choose a machine above to see a verdict for each format. Method
GGUF for llama.cpp, Ollama, LM Studio and others
unsloth/Inkling-Small-GGUF (converted by unsloth), 137k downloads. Sizes are the real files on the Hub, all parts added up. For image input you also need the projector file mmproj-BF16.gguf (0.13 GB), not counted below.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| UD-IQ3_XXS | 91.2 GBfile size, 3 parts | 93.0 GB91.2 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-IQ3_XXSDownload hf download unsloth/Inkling-Small-GGUF --include "UD-IQ3_XXS/Inkling-Small-UD-IQ3_XXS-*-of-00003.gguf" --local-dir Inkling-Small-GGUF |
| UD-IQ3_S | 100.1 GBfile size, 4 parts | 101.9 GB100.1 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-IQ3_SDownload hf download unsloth/Inkling-Small-GGUF --include "UD-IQ3_S/Inkling-Small-UD-IQ3_S-*-of-00004.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q3_K_M | 111.2 GBfile size, 4 parts | 113.0 GB111.2 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q3_K_MDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q3_K_M/Inkling-Small-UD-Q3_K_M-*-of-00004.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q3_K_XL | 111.3 GBfile size, 4 parts | 113.2 GB111.3 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q3_K_XLDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q3_K_XL/Inkling-Small-UD-Q3_K_XL-*-of-00004.gguf" --local-dir Inkling-Small-GGUF |
| UD-IQ4_XS | 118.7 GBfile size, 4 parts | 120.5 GB118.7 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-IQ4_XSDownload hf download unsloth/Inkling-Small-GGUF --include "UD-IQ4_XS/Inkling-Small-UD-IQ4_XS-*-of-00004.gguf" --local-dir Inkling-Small-GGUF |
| UD-IQ4_NL | 121.2 GBfile size, 4 parts | 123.0 GB121.2 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-IQ4_NLDownload hf download unsloth/Inkling-Small-GGUF --include "UD-IQ4_NL/Inkling-Small-UD-IQ4_NL-*-of-00004.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q4_K_S | 141.9 GBfile size, 5 parts | 143.7 GB141.9 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q4_K_SDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q4_K_S/Inkling-Small-UD-Q4_K_S-*-of-00005.gguf" --local-dir Inkling-Small-GGUF |
| MXFP4_MOE | 147.2 GBfile size, 5 parts | 149.0 GB147.2 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:MXFP4_MOEDownload hf download unsloth/Inkling-Small-GGUF --include "MXFP4_MOE/Inkling-Small-MXFP4_MOE-*-of-00005.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q4_K_M | 151.4 GBfile size, 5 parts | 153.2 GB151.4 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q4_K_MDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q4_K_M/Inkling-Small-UD-Q4_K_M-*-of-00005.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q4_K_XL | 152.1 GBfile size, 5 parts | 153.9 GB152.1 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q4_K_XLDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q4_K_XL/Inkling-Small-UD-Q4_K_XL-*-of-00005.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q5_K_S | 171.6 GBfile size, 5 parts | 173.4 GB171.6 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q5_K_SDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q5_K_S/Inkling-Small-UD-Q5_K_S-*-of-00005.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q5_K_M | 182.6 GBfile size, 5 parts | 184.4 GB182.6 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q5_K_MDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q5_K_M/Inkling-Small-UD-Q5_K_M-*-of-00005.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q5_K_XL | 183.2 GBfile size, 5 parts | 185.0 GB183.2 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q5_K_XLDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q5_K_XL/Inkling-Small-UD-Q5_K_XL-*-of-00005.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q6_K | 203.9 GBfile size, 6 parts | 205.7 GB203.9 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q6_KDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q6_K/Inkling-Small-UD-Q6_K-*-of-00006.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q6_K_XL | 223.2 GBfile size, 6 parts | 225.1 GB223.2 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q6_K_XLDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q6_K_XL/Inkling-Small-UD-Q6_K_XL-*-of-00006.gguf" --local-dir Inkling-Small-GGUF |
| Q8_0 | 261.0 GBfile size, 7 parts | 262.8 GB261.0 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:Q8_0Download hf download unsloth/Inkling-Small-GGUF --include "Q8_0/Inkling-Small-Q8_0-*-of-00007.gguf" --local-dir Inkling-Small-GGUF |
| UD-Q8_K_XL | 268.6 GBfile size, 7 parts | 270.4 GB268.6 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:UD-Q8_K_XLDownload hf download unsloth/Inkling-Small-GGUF --include "UD-Q8_K_XL/Inkling-Small-UD-Q8_K_XL-*-of-00007.gguf" --local-dir Inkling-Small-GGUF |
| BF16 | 491.2 GBfile size, 11 parts | 493.0 GB491.2 + 1.3 + 0.50 | Commandsllama.cpp llama-server -hf unsloth/Inkling-Small-GGUF:BF16Download hf download unsloth/Inkling-Small-GGUF --include "BF16/Inkling-Small-BF16-*-of-00011.gguf" --local-dir Inkling-Small-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLX 4bitpipenetwork/Inkling-Small-MLX-REAP25-4bit | 104.4 GBfile size | 106.3 GB104.4 + 1.3 + 0.50 | Download page |
| MLX mxfp4mlx-community/Inkling-Small-mxfp4 | 130.6 GBfile size | 132.5 GB130.6 + 1.3 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| BF16original weights | 495.4 GBfile size | 497.2 GB495.4 + 1.3 + 0.50 |
hf download thinkingmachines/Inkling-Small --local-dir Inkling-SmallOther conversions (16; sizes not read)
- thinkingmachines/Inkling-Small-NVFP4 NVFP4, by the maker, 82k downloads
- xautonomics/Inkling-Small-UD-Q4_K_M.LAYR.GGUF GGUF, 1.1k downloads
- RedHatAI/Inkling-Small-FP8-dynamic FP8, 481 downloads
- intel-ai/Inkling-Small-MXFP4-AutoRound Other, mxfp4, 174 downloads
- EmbeddedLLM/Inkling-Small-MXFP4 Other, mxfp4, 149 downloads
- cyankiwi/Inkling-Small-AWQ-INT4 AWQ, int4, 126 downloads
- AxionML/Inkling-Small-NVFP4 NVFP4, 73 downloads
- takumar24/Inkling-Small-FP8-dynamic FP8, 53 downloads
- 0xSero/Inkling-Small-EXL3-2.0bpw EXL3, 47 downloads
- INCModel2/Inkling-Small-MXFP4-CT-AutoRound Other, mxfp4, 44 downloads
- olka-fi/Inkling-Small-MXFP4 Other, mxfp4, 41 downloads
- eyes-dot-ml/Inkling-Small-NVFP4 NVFP4, 36 downloads
- konizquants/Inkling-Small-NVFP4 NVFP4, 36 downloads
- 0xSero/Inkling-Small-EXL3-2.5bpw EXL3, 26 downloads
- 0xSero/Inkling-Small-EXL3-3.5bpw EXL3, 22 downloads
- 0xSero/Inkling-Small-EXL3-4.5bpw EXL3, 22 downloads
The Hub tags 55 quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.
Run it by API
1 provider serves this model through the Hugging Face router. Prices are US dollars per million tokens. Precision is what the provider reports; a lower precision than the original weights can change answers. Live first, then cheapest output.
| Provider | In | Out | Context | Precision | Tokens/s | First token | Uptime 24h | Tools | JSON |
|---|---|---|---|---|---|---|---|---|---|
| DeepInfra | $0.45 | $1.20 | 512k | fp4 | 121 | 0.4 s | 100.0% | yes | no |
Through the router, with a Hugging Face token (the router bills your Hugging Face account). Change the name after the colon to pick another provider.
curl https://router.huggingface.co/v1/chat/completions \
-H "Authorization: Bearer $HF_TOKEN" -H "Content-Type: application/json" \
-d '{"model": "thinkingmachines/Inkling-Small:deepinfra", "messages": [{"role": "user", "content": "Hello"}]}'Not built yet: one key for every open model, with BlockRun, and our own tested-quality score per provider. Until then, precision and measured speed are the best guide to a provider here.
Planned for this page Not built yet
- Try it in private chat
- A link to Solana AI's confidential chat for the models it serves. This model is not listed as served there.
- Fine-tune it
- Fine-tuning a few base models from this page and publishing the result here.
- Creators and earnings
- Pages for people who publish fine-tunes and adapters of this model, and what they earn.
- Fund and bounties
- Bounties for new open models, paid from a public fund. How the fund will work.
- Contribute compute
- How to give compute to open model work, with Akash.
Model data from the Hugging Face Hub 19 Nov 2026, 13:43 UTC. The repo. Found a mistake on this page? Tell us.