qwen3-vl-8b-hebrew-merged by Isaac Godfried: Open Model

Runs On

What it takes to serve qwen3-vl-8b-hebrew-merged (8.8B 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	17.5 GB	21.0 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
8-bit	8.8 GB	10.5 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
4-bit	4.4 GB	5.3 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x 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, read Oct 12, 2026.

qwen3-vl-8b-hebrew-merged on every accelerator the SAVRN Index prices, at every precision

Model Card

By Isaac Godfried, published under apache-2.0, revision 2cbbb8cfef2a.

This is a historical merged export, NOT the current flagship. The flagship is v1.9a — adapter + merge instructions at qwen3-vl-8b-hebrew-v19a-ckpt. Full 16-bit merge of the series' first-generation (v1) Talmud adapter — the first model in this line to read Vilna-print gemara from pixels. Vision tower unadapted (language-only LoRA generation). Superseded on every (rope under textconfig.ropeparameters); MLX conversion requires patching ropetheta to top level. One line per generation — which checkpoint to use and which are historical: \\ CER over substantive attempts only, alignment-based scorer. Benchmarks differ across generations (Talmud print vs Genizah manuscripts) — compare within a…

Read Isaac Godfried's full model card

Qwen3-VL-8B Hebrew — v1 merged (historical)

Cairo Genizah AI Transcription — formal analysis and write-up of the project The full account of the methods, training data, benchmarks and error analysis behind the Cairo Genizah AI transcription models. Search the transcribed corpus at Cairo Genizah AI.

This is a historical merged export, NOT the current flagship. The flagship isv1.9a — adapter + merge instructions at qwen3-vl-8b-hebrew-v19a-ckpt.

Full 16-bit merge of the series' first-generation (v1) Talmud adapter — the first model in this line to read Vilna-print gemara from pixels. Vision tower unadapted (language-only LoRA generation). Superseded on every benchmark by later versions. Note: the saved config is transformers-5-style (rope under text_config.rope_parameters); MLX conversion requires patching rope_theta to top level.

The series at a glance

One line per generation — which checkpoint to use and which are historical:

version	repo	status	canonical revision	headline (benchmark)
v1.9a	v19a-ckpt	FLAGSHIP — use this	step 1300 `43e21bd7`	F1 0.816 / CER 0.216 (Genizah religious-140); F1 0.862 / CER 0.196 (frozen PGP-131)
v1.9b	v19b-ckpt	experimental control — not adopted	step 1300 `f8618c28`	merger-LoRA ablation study; loss ≡ v1.9a
v1.8b	v18b-ckpt	superseded; warm-start ancestor of all v1.9	step 700 `c80313f8`	first arm with a live vision-tower LoRA
v1.8a	v18a-ckpt	superseded A/B control (vision frozen)	step 700	control arm for the v1.8 vision experiment
v1.7	v17-ckpt	superseded	step 800	first Genizah-handwriting generation; best VLM on both corpora at its era's benchmarks (Aug 2026)
v1.6	v16-ckpt	superseded; Talmud-print reference	step 1000 (official); step 1100 `b9f47f32` (v1.7 warm start)	Talmud page CER: gemara 0.090 / rashi 0.047 / tosafot 0.099 — Rashi-script 6.7× better than the best closed model we tested (0.315)
v1.5	rashi-ckpt	superseded	final `e050aca8` (step 2000)	proved pure Rashi-glyph perception: CER 0.018 on unmemorizable synthetic text
v1	hebrew-ckpt	superseded (language-only LoRA)	step 3800 `e2f85dde`	first checkpoints to read Vilna gemara from pixels

\* CER over substantive attempts only, alignment-based scorer. Benchmarks differ across generations (Talmud print vs Genizah manuscripts) — compare within a row's named benchmark, not across rows.

Configuration

Architecture: Qwen3VLForConditionalGeneration
Context length (tokens): 262,144
Layers: 36
Hidden size: 4,096
Feed-forward size: 12,288
Attention heads: 32
Key/value heads: 8
Head dimension: 128
Vocabulary size: 151,936
Stored precision: bfloat16
Model type: qwen3_vl

Identity and Version

Repository: isaacmg/qwen3-vl-8b-hebrew-merged
Publisher: Isaac Godfried
Task: Image and text to text
Modality: Image and text
Library: transformers
Parameters: 8.8B parameters
Languages: he
Revision: 2cbbb8cfef2a1abcee95ecef02623f49ac9c4764
First published: 2026-07-17
Last updated: 2026-10-05

Files and Weights

13 files, 17.5 GB in total. The weights are 4 files totalling 17.5 GB in safetensors.

Weights4 files · 17.5 GB

Configuration4 files · 71.3 KB

Tokenizer2 files · 11.4 MB

Documentation1 file · 3.5 KB

Other1 file · 5.3 KB

Repository1 file · 1.6 KB

Every file

File	Type	Size	SHA-256
model-00001-of-00004.safetensors	Weights	4.9 GB	5a9d75f10d2f
model-00002-of-00004.safetensors	Weights	4.9 GB	485c92550602
model-00003-of-00004.safetensors	Weights	5.0 GB	f1ef8426f7b0
model-00004-of-00004.safetensors	Weights	2.7 GB	9c31af7e40fc
config.json	Configuration	2.0 KB	—
generation_config.json	Configuration	186 B	—
model.safetensors.index.json	Configuration	67.8 KB	—
processor_config.json	Configuration	1.3 KB	—
README.md	Documentation	3.5 KB	—
chat_template.jinja	Other	5.3 KB	—
.gitattributes	Repository	1.6 KB	—
tokenizer.json	Tokenizer	11.4 MB	afbeda012dca
tokenizer_config.json	Tokenizer	10.6 KB	—

License and Download

License: apache-2.0
Access: Open weights, no gate
Download size: 17.5 GB

Download from Isaac Godfried

Released by Isaac Godfried through its official repository on Hugging Face. Read the license.

Built From

Derived from Qwen/Qwen3-VL-8B-Instruct

Memory Requirements

Precision	Weights in memory
As published	17.5 GB
16-bit	17.5 GB
8-bit	8.8 GB
4-bit	4.4 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About qwen3-vl-8b-hebrew-merged

How much GPU memory does qwen3-vl-8b-hebrew-merged need?

About 21 GB at 16-bit and 5.3 GB at 4-bit: the weights (8.8B parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run qwen3-vl-8b-hebrew-merged 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 qwen3-vl-8b-hebrew-merged commercially?

Yes. qwen3-vl-8b-hebrew-merged 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 qwen3-vl-8b-hebrew-merged's context length?

262,144 tokens, from the maximum position embeddings in its published configuration.

Similar Models

Model · Image and text to text

Qwen3-VL-8B-Instruct

Qwen

Meet Qwen3-VL — the most powerful vision-language model in the Qwen series to date. This generation delivers comprehensive upgrades across the board: superior text understanding & generation, deeper visual perception & reasoning, extended context length, enhanced spatial and video dynamics comprehension, and stronger agent interaction capabilities. Available in Dense and MoE architectures that scale from edge to cloud, with Instruct and reasoning‑enhanced Thinking editions for flexible, on‑demand deployment. Text Understanding on par with pure LLMs: Seamless text–vision fusion for lossless, unified comprehension. 1. Interleaved-MRoPE: Full‑frequency allocation over time, width, and height…

Open weights apache-2.0 8.8B parameters 262,144 tokens transformers

View model

Model · Image and text to text

qwen3-vl-8b-hebrew-v20a-merged

Isaac Godfried

Merged bf16 weights of the v2.0a checkpoint (v20a-ckpt, step 1800, commit af9df6a0): the LoRA adapters (language layers r16, vision tower) folded into Qwen3-VL-8B-Instruct. Ready for transformers, vLLM, or conversion. The image processor ships the model's resolution contract (minpixels=6500000, maxpixels=7000000): send the largest image you have and let the processor resize it. Status: current best of the series on both transcription benchmarks (Sep 2026) and the first checkpoint with text grounding. Serving demo: the Genizah Reader Space. Diplomatic transcription of handwritten Hebrew-script manuscripts (Hebrew, Aramaic, Judaeo-Arabic) — Cairo Genizah documentary fragments and…

Open weights apache-2.0 8.8B parameters 262,144 tokens transformers

View model

Model · Image and text to text

qwen3-vl-8b-hebrew-rashi-merged

Isaac Godfried

This is a historical merged export, NOT the current flagship. The flagship is v1.9a — adapter + merge instructions at qwen3-vl-8b-hebrew-v19a-ckpt. Full 16-bit merge of v1.5, the Rashi-script generation: pure Rashi glyph perception proven (CER 0.018 on unmemorizable synthetic text), crop-level Vilna gemara solved (CER 0.019), Talmud page-level gemara CER 0.059. Trained at a 3.2 MP resolution floor — use that policy at inference. Vision tower unadapted. Superseded by v1.6+ on print and v1.9a on manuscripts. One line per generation — which checkpoint to use and which are historical: \\ CER over substantive attempts only, alignment-based scorer. Benchmarks differ across generations (Talmud…

Open weights apache-2.0 8.8B parameters 262,144 tokens transformers

View model

Model · Image and text to text

MemGUI-8B-RL

Anonymous

Anonymous release for the ICLR 2027 submission MemGUI-RL: Reinforcement Learning for Proactive Context Management in Long-Horizon Mobile GUI Agents. Project page: https://memgui-rl-anonymous.github.io/ MemGUI-8B-RL is MemGUI-8B-SFT (Qwen3-VL-8B-Instruct supervised on MemGUI-3K) post-trained for 100 optimizer steps with FARPO (Folding-Aware Reward-decoupled Policy Optimization, span-to-step ratio rho = 9). The policy speaks the ConAct (Context-as-Action) interface of MemGUI-Agent: every response contains a folding directive for its own history, an optional memory operation and the next GUI action. The checkpoint is a standard Qwen3VLForConditionalGeneration model (weights in bf16, ~17.5 GB).…

Open weights apache-2.0 8.8B parameters 262,144 tokens

View model

Model · Image and text to text

Legend_ocr_qwen3-highavg

Jackson Kahungu

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 8.8B parameters 262,144 tokens transformers

View model

Model · Image and text to text

Qwen3-VL-8B-Instruct-FP8

Qwen

Meet Qwen3-VL — the most powerful vision-language model in the Qwen series to date. This generation delivers comprehensive upgrades across the board: superior text understanding & generation, deeper visual perception & reasoning, extended context length, enhanced spatial and video dynamics comprehension, and stronger agent interaction capabilities. Available in Dense and MoE architectures that scale from edge to cloud, with Instruct and reasoning‑enhanced Thinking editions for flexible, on‑demand deployment. Text Understanding on par with pure LLMs: Seamless text–vision fusion for lossless, unified comprehension. 1. Interleaved-MRoPE: Full‑frequency allocation over time, width, and height…

Open weights apache-2.0 8.8B parameters 262,144 tokens transformers

View model