Josev-4B has 4.238 billion total parameters. The selected Q8_0 GGUF file is 4.517 GB.
Data snapshot: . 1 published test in this catalog.
Model file and details
Provider
Mayflower GmbH
Total parameters
4,237,797,892
Exact file size
4,516,653,568 bytes
Quantization / precision
Q8_0
Weight repository
mayflowergmbh/Josev-4B-GGUF
Original model repository
mayflowergmbh/Josev-4B
Weight revision
d60a12766d1f03e4930a1d66aa0cca213080341a
Model revision
74270d432d6d4ac86f1571766e991662e52a0877
Runtime notes
Confirm model architecture and quantization support in your runner.
Recorded release date
2026-10-15
Date evidence
Mayflower GmbH released the Apache-2.0 Josev-4B source checkpoint and official same-publisher GGUF set on 15 October 2026. The selected 4,516,653,568-byte Q8_0 language file is the publisher's recommended GGUF and was measured against the bf16 source model on the decisions-de validation split. Josev is a German specialist for typed classification decisions rather than a general chat model; image and video input additionally require a separate 366,894,144-byte Q8_0 projector.
These scores were compiled from the linked sources. These published scores were not independently tested here. Different prompts, precision, thinking budgets and tools can change results.
Josev-4B Q8_0 · decisions-de validation
85.8 / 100
Publisher-reported exact-decision accuracy for the selected Josev-4B Q8_0 GGUF on the German decisions-de validation split. This is a specialist in-domain classification test, not a general chat benchmark.
Benchmark efficiency: 18.996 published score points per GB of selected weights. This ratio is not an inference-speed measure; check the test settings to see whether the score covers the selected quantization.
Test settings: Publisher-run 1,215-decision accuracy on 565 cases from the decisions-de validation split using the exact Josev-4B Q8_0 GGUF through llama-server /v1/systemone. The card compares it with the bf16 PyTorch source checkpoint under the same question-type temperatures; Q8_0 scored 85.8% versus 85.9% and matched 78.1% whole-case accuracy. The export requires llama.cpp b11443 or newer and was tested with b11541.
Publisher measurement, not independent validation. decisions-de is the in-domain German training family and combines 90 task sources; the source model was trained on its train split, so this validation score is not evidence of general chat quality or performance on unseen domains. Calibration temperatures were fitted on this same validation split, making the reported ECE unsuitable as independent calibration evidence. Q4_K_M and Q6_K were not measured, and the publisher warns that lower quantizations can shift probabilities and flip close decisions. The selected 4.52 GB language file excludes the optional 366.9 MB Q8_0 vision projector, runtime, context and batch headroom. llama-server batch and micro-batch sizes must cover the full request or longer requests fail; file size alone does not establish compatibility or speed on a computer. Apache-2.0 license.
Context and version evidence: The repository and file name identify the selected model. Revision hashes are shown above when recorded. Tested context length, full hardware configuration and run date are not consistently recorded in this snapshot. Only settings explicitly stated in each benchmark record are known.
The 4.517 GB figure covers model weights in this file. It excludes the context cache, runtime, operating system and any separate multimodal files. It is not a measured RAM requirement or a token-generation speed.
These complete configurations pass the initial weight-size check for the selected 4.52 GB file. Allow extra memory for context and software. Check the model’s runtime notes and your inference software before buying.
Listing evidence: 2026-09-25. Amazon.com base price visible in an owner-supplied screenshot reviewed 25 Sep 2026; capture time not supplied. Dated reference, not a live quote. Excludes card signup offers, coupons, tax and shipping.
Listing evidence: 2026-10-03. Apple US launch price announced 22 September 2026: USD 1,699. Dated manufacturer reference, not an Amazon offer; tax and shipping may differ.
Listing evidence: 2026-09-24. GMKtec, EVO-X2 / 64GB / 1TB, recorded 22 Sep 2026; exact ASIN capacity identified 24 Sep. This is a dated reference, not a current Amazon price; tax, shipping and bundle contents may differ.
Memory matches are estimates, not tested compatibility or speed. As an Amazon Associate I earn from qualifying purchases.
Confirm model architecture and quantization support in your runner. The selected weights use 4.517 GB; this excludes context cache, runtime buffers and other programs.
These are the first listed tiers that pass our reserve policy, not measured minimum requirements. System RAM and VRAM remain separate. Use the memory tool to change the reserve for your context and runner.