# StandardOne-8B by Standard Thinking: Open-Weight Model Source: https://savrn.com/models/standardone-8b Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve StandardOne-8B (8.9B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 17.8 GB | 21.4 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 8.9 GB | 10.7 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 4.5 GB | 5.4 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 7, 2026. [StandardOne-8B on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/standardone-8b/gpus) ## Model Card By Standard Thinking, published under apache-2.0, revision d8ef2dc6e131. Updated weights (v2, 2026-09-26). If you downloaded this model before, download it again or pin revision="v2". Earlier versions stay available under the tags v1 and v1.1. Version: v2 Standard One scores a bounded set of answers for a supplied scenario and returns probabilities through POST /v1/systemone. It does not generate free-form response text. This repository contains the merged BF16 8B checkpoint and the server code. | If you need | Repository | | --- | --- | | Merged 8B checkpoint and server code | [StandardOne-8B](https://savrn.com/models/standardone-8b) (this repository) | | 8B adapter weights and merge recipe | [StandardOne-8B-LoRA](https://savrn.com/models/standardone-8b-lora) | | Smaller merged checkpoint | [StandardOne-3B](https://savrn.com/models/standardone-3b) | | Smaller adapter weights and merge recipe | [StandardOne-3B-LoRA](https://savrn.com/models/standardone-3b-lora) | In the reported served evaluations, 8B scores higher than 3B on the public standard and hard tiers; 3B has a lower median latency on the measured short-request profile. See Benchmarks for the measurement conditions and limitations. The figure combines results from different measurement paths. See Benchmarks for served versus offline conditions; measured 24–26 September 2026. ### At a glance [Read the full model card (2,026 words)](https://savrn.com/models/standardone-8b/card) ## Configuration Architecture Mistral3ForConditionalGeneration Context length (tokens) 262,144 Layers 34 Hidden size 4,096 Feed-forward size 14,336 Attention heads 32 Key/value heads 8 Head dimension 128 Vocabulary size 131,072 Model type mistral3 ## Identity and Version Repository StandardThinking/StandardOne-8B Publisher Standard Thinking Task Text generation Modality Text Library transformers Parameters 8.9B parameters Languages en, ja, zh, es, fr, de, pt, ru Revision d8ef2dc6e1316f7ebc7e9d7a313e65cb7a2588bd First published 2026-09-24 Last updated 2026-09-27 ## Files and Weights 94 files, 17.9 GB in total. The weights are 4 files totalling 17.8 GB in safetensors. Weights4 files · 17.8 GB Configuration48 files · 17.3 MB Tokenizer4 files · 17.3 MB Documentation19 files · 141.0 KB Other17 files · 2.0 MB Repository2 files · 2.0 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model-00001-of-00004.safetensors | Weights | 5.0 GB | 5b29a2e3e863 | | model-00002-of-00004.safetensors | Weights | 5.0 GB | 65c9000a4aaf | | model-00003-of-00004.safetensors | Weights | 4.9 GB | fe537b22d9b0 | | model-00004-of-00004.safetensors | Weights | 2.9 GB | 6c3f84685c69 | | MERGE_REPORT.json | Configuration | 2.7 KB | — | | config.json | Configuration | 1.7 KB | — | | evidence/calibration-leaderboard-ablate-nosys.json | Configuration | 21.8 KB | — | | evidence/calibration-leaderboard-merged-native.json | Configuration | 10.9 KB | — | | evidence/calibration-per-type-served.json | Configuration | 14.1 KB | — | | evidence/prompt-wording-choice.json | Configuration | 2.3 KB | — | | evidence/served-nosys-easy-report.json | Configuration | 6.9 KB | — | | evidence/served-nosys-hard-report.json | Configuration | 14.8 KB | — | | evidence/served-nosys-latency-idle-gpu.json | Configuration | 43.6 KB | — | | evidence/served-nosys-original-report.json | Configuration | 9.6 KB | — | | generation_config.json | Configuration | 131 B | — | | model.safetensors.index.json | Configuration | 52.7 KB | — | | params.json | Configuration | 1.1 KB | — | | processor_config.json | Configuration | 976 B | — | | release-manifest.json | Configuration | 1.1 KB | — | | server/benchmarks/data/jevbench-easy/public.manifest.json | Configuration | 3.3 KB | — | | server/benchmarks/data/jevbench-hard/public.manifest.json | Configuration | 3.7 KB | — | | server/benchmarks/data/jevbench-original/public.manifest.json | Configuration | 3.4 KB | — | | server/benchmarks/run_matrix.py | Configuration | 9.4 KB | — | | server/benchmarks/verify_metrics.py | Configuration | 5.1 KB | — | | server/examples/request.json | Configuration | 542 B | — | | server/examples/smoke.py | Configuration | 1.9 KB | — | | server/jev_adapter/__init__.py | Configuration | 73 B | — | | server/jev_adapter/__main__.py | Configuration | 7.5 KB | — | | server/jev_adapter/backend.py | Configuration | 892 B | — | | server/jev_adapter/benchmarks/__init__.py | Configuration | 81 B | — | | server/jev_adapter/benchmarks/compare.py | Configuration | 2.8 KB | — | | server/jev_adapter/benchmarks/data.py | Configuration | 8.5 KB | — | | server/jev_adapter/benchmarks/jevbench.py | Configuration | 12.1 KB | — | | server/jev_adapter/benchmarks/metrics.py | Configuration | 12.1 KB | — | | server/jev_adapter/benchmarks/prepare.py | Configuration | 8.3 KB | — | | server/jev_adapter/benchmarks/run.py | Configuration | 18.4 KB | — | | server/jev_adapter/protocol.py | Configuration | 15.6 KB | — | | server/jev_adapter/server.py | Configuration | 4.4 KB | — | | server/jev_adapter/service.py | Configuration | 7.0 KB | — | | server/jev_adapter/sglang.py | Configuration | 19.5 KB | — | | server/tests/test_benchmark_data.py | Configuration | 8.7 KB | — | | server/tests/test_benchmark_matrix.py | Configuration | 3.4 KB | — | | server/tests/test_benchmark_run.py | Configuration | 10.0 KB | — | | server/tests/test_disconnect_cleanup.py | Configuration | 5.1 KB | — | | server/tests/test_http_integration.py | Configuration | 2.0 KB | — | | server/tests/test_jevbench.py | Configuration | 11.0 KB | — | | server/tests/test_main.py | Configuration | 11.5 KB | — | | server/tests/test_protocol.py | Configuration | 17.8 KB | — | | server/tests/test_service.py | Configuration | 17.1 KB | — | | server/tests/test_sglang.py | Configuration | 16.0 KB | — | | special_tokens_map.json | Configuration | 147.1 KB | — | | tekken.json | Configuration | 16.8 MB | 600bb2794656 | | LICENSE | Documentation | 11.3 KB | — | | NOTICE | Documentation | 1.2 KB | — | | QUICKSTART.md | Documentation | 6.1 KB | — | | README.md | Documentation | 16.8 KB | — | | docs/BENCHMARKS.md | Documentation | 26.0 KB | — | | docs/public-classification-suites.md | Documentation | 3.3 KB | — | | server/LICENSE | Documentation | 11.3 KB | — | | server/NOTICE | Documentation | 931 B | — | | server/README.md | Documentation | 17.5 KB | — | | server/benchmarks/DECISION_BENCHMARK_SELECTION.md | Documentation | 5.2 KB | — | | server/benchmarks/METRIC_VALIDATION.md | Documentation | 3.8 KB | — | | server/benchmarks/PUBLIC_DATASETS.md | Documentation | 13.4 KB | — | | server/benchmarks/README.md | Documentation | 12.8 KB | — | | server/benchmarks/data/jevbench-easy/LICENSE | Documentation | 1.1 KB | — | | server/benchmarks/data/jevbench-easy/THIRD-PARTY.md | Documentation | 2.7 KB | — | | server/benchmarks/data/jevbench-hard/LICENSE | Documentation | 1.1 KB | — | | server/benchmarks/data/jevbench-hard/THIRD-PARTY.md | Documentation | 2.7 KB | — | | server/benchmarks/data/jevbench-original/LICENSE | Documentation | 1.1 KB | — | | server/benchmarks/data/jevbench-original/THIRD-PARTY.md | Documentation | 2.7 KB | — | | SHA256SUMS | Other | 9.0 KB | — | | SYSTEM_PROMPT.txt | Other | 2.4 KB | — | | chat_template.jinja | Other | 11.9 KB | — | | docs/assets/00-benchmark-card.png | Other | 297.2 KB | e7e9f42a23bf | | docs/assets/00-benchmark-card.svg | Other | 58.7 KB | — | | docs/assets/02-latency-vs-qwen.png | Other | 114.5 KB | 7313f2b0645e | | docs/assets/02-latency-vs-qwen.svg | Other | 11.3 KB | — | | docs/assets/04-throughput.png | Other | 97.7 KB | — | | docs/assets/04-throughput.svg | Other | 8.7 KB | — | | docs/assets/05-gain-over-base.png | Other | 179.2 KB | 95a53a990799 | | docs/assets/05-gain-over-base.svg | Other | 16.6 KB | — | | docs/assets/06-vs-jev.png | Other | 267.8 KB | c43fb809f0b6 | | docs/assets/06-vs-jev.svg | Other | 23.0 KB | — | | server/benchmarks/data/jevbench-easy/public.jsonl | Other | 60.7 KB | — | | server/benchmarks/data/jevbench-hard/public.jsonl | Other | 708.2 KB | — | | server/benchmarks/data/jevbench-original/public.jsonl | Other | 91.9 KB | — | | server/pyproject.toml | Other | 808 B | — | | .gitattributes | Repository | 1.9 KB | — | | server/.gitignore | Repository | 125 B | — | | server/jev_adapter/native_tokenizer.py | Tokenizer | 7.5 KB | — | | server/tests/test_native_tokenizer.py | Tokenizer | 14.1 KB | — | | tokenizer.json | Tokenizer | 17.1 MB | d5f6046775b1 | | tokenizer_config.json | Tokenizer | 198.1 KB | — | ## License and Download License apache-2.0 Access Open weights, no gate Download size 17.8 GB [Download from Standard Thinking](https://huggingface.co/StandardThinking/StandardOne-8B) Released by Standard Thinking through its official repository on Hugging Face. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Built From - Derived from mistralai/Ministral-3-8B-Instruct-2512-BF16 ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 17.8 GB | | 16-bit | 17.8 GB | | 8-bit | 8.9 GB | | 4-bit | 4.5 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Built on This Model - Quantized from[StandardOne-8B-GGUF](https://savrn.com/models/standardone-8b-gguf) - Derived from[StandardOne-8B-GGUF](https://savrn.com/models/standardone-8b-gguf) - Quantized from[StandardOne-8B-FP8](https://savrn.com/models/standardone-8b-fp8) - Derived from[StandardOne-8B-FP8](https://savrn.com/models/standardone-8b-fp8) ## Questions About StandardOne-8B ### How much GPU memory does StandardOne-8B need? About 21.4 GB at 16-bit and 5.4 GB at 4-bit: the weights (8.9B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run StandardOne-8B on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### Can I use StandardOne-8B commercially? Yes. StandardOne-8B is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors. ### What is StandardOne-8B's context length? 262,144 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Text generation ### [StandardOne-8B-FP8](https://savrn.com/models/standardone-8b-fp8) [Standard Thinking](https://savrn.com/model-publishers/standardthinking) StandardOne-8B-FP8 is an FP8 (compressed-tensors, float8e4m3 weights, dynamic per-token activations) quantization of the released StandardOne-8B decision model. The language-model linear projections (q/k/v/o, gate/up/down) are quantized per-channel FP8 E4M3 with dynamic FP8 activations (llm-compressor's data-free FP8DYNAMIC recipe, no calibration data required); the vision tower, multi-modal projector, embeddings and lmhead are left unquantized in BF16. It was produced from source revision e88423700bb5ab9b2f50e176cf19825914345272 of StandardOne-8B on 2026-09-25 using llm-compressor 0.14.0 (torch 2.14.0, transformers 5.17.0, compressed-tensors 0.19.0); results below. Served through SGLang… Open weights apache-2.0 8.9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/standardone-8b-fp8) Model · Text generation ### [idefics-9b](https://savrn.com/models/idefics-9b) [HuggingFaceM4](https://savrn.com/model-publishers/huggingfacem4) How do I pronounce the model's name? Watch a Youtube tutorial IDEFICS (Image-aware Decoder Enhanced à la Flamingo with Interleaved Cross-attentionS) is an open-access reproduction of Flamingo, a closed-source visual language model developed by Deepmind. Like GPT-4, the multimodal model accepts arbitrary sequences of image and text inputs and produces text outputs. IDEFICS is built solely on publicly available data and models. The model can answer questions about images, describe visual contents, create stories grounded on multiple images, or simply behave as a pure language model without visual inputs. IDEFICS is on par with the original closed-source model on various image-text benchmarks… Open weights other 8.9B parameters 2,048 tokens transformers [View model](https://savrn.com/models/idefics-9b) Model · Text generation ### [dQwen3.5-9B-Base](https://savrn.com/models/dqwen3-5-9b-base) [IFML](https://savrn.com/model-publishers/ut-ifml) A masked diffusion language model adapted from Qwen3.5-9B. The backbone is hybrid: only its attention layers are made bidirectional, and the Gated DeltaNet layers stay causal. This is a base model, with no instruction tuning. Paper: dQwen3.5: Hybrid-Attention Diffusion Language Models. Code: https://github.com/AntonXue/dQwen Needs a CUDA GPU and transformers>=5.13 (tested with torch 2.7.1+cu128, flash-linear-attention 0.5.1). generate decodes the whole canvas at once, committing positions above a confidence threshold (tau=0.9); pass blocklength=32 for left-to-right block decoding, or tau=None, stepsperblock=k for a fixed budget. The 50B-token checkpoint from the paper is… Open weights apache-2.0 9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/dqwen3-5-9b-base) Model · Text generation ### [nexgen-flash-v1-merged-averaged](https://savrn.com/models/nexgen-flash-v1-merged-averaged) [Sampson Kontomah](https://savrn.com/model-publishers/nexgen2) basemodel: - Qwen/Qwen3.5-9B pipelinetag: text-generation libraryname: transformers Open weights apache-2.0 9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/nexgen-flash-v1-merged-averaged) Model · Text generation ### [WikiQwen-9B](https://savrn.com/models/wikiqwen-9b) [Devon Yanitski](https://savrn.com/model-publishers/devon7y) Ask anything, get a how-to article. WikiQwen-9B is Qwen3.5-9B fine-tuned to answer every message the same way: as a tidy, step-by-step how-to article in markdown, with a picture caption above each step. Hand those captions to WikiQwen-Illustrator and they become illustrations. Send it a question ("how do I keep basil alive?"), a problem ("my bike chain keeps falling off") or just "hi", and it sends back a guide. Every reply has the same shape: - a # How to … title and a short intro - one or more ## Method N: or ## Part N: sections, or a single ## Steps - an [IMAGE: caption] line above each step, written for the Illustrator to draw - steps written as N. Bold summary. Details…, with bullets… Open weights cc-by-nc-sa-4.0 9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/wikiqwen-9b) Model · Text generation ### [zkhapo3.5-9b](https://savrn.com/models/zkhapo3-5-9b) [Zkhapo](https://savrn.com/model-publishers/zkhapo) This is the model card of a transformers model that has been pushed on the Hub. Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). Open weights 9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/zkhapo3-5-9b) ## Standard Thinking [All models and datasets](https://savrn.com/model-publishers/standardthinking) ## Versions - [d8ef2dc6e131](https://savrn.com/models/standardone-8b/versions/d8ef2dc6e131) · current 2026-09-27 - [e88423700bb5](https://savrn.com/models/standardone-8b/versions/e88423700bb5) 2026-09-25 ## Explore More - [All text generation models](https://savrn.com/models/tasks/text-generation) - [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-09-27. - [Hugging Face record](https://huggingface.co/StandardThinking/StandardOne-8B) - [How the hub is built](https://savrn.com/model-hub/methodology)