Up to date
kirp CC BY-NC 4.0 out 31 days ago
4.54B parameters, images and text in, text out, 256K tokens of context.
Verified 2026-10-25 Rank #0 in Decision models · score 0 method v1 Sources 3 Downloads 30d 1.5k · #584 of 619 tracked
Picked by rule from 1 tracked piece: the lab's own post, the most shared, a field report, the latest.
SPE, Aramco Americas, and i2k Connect Inc. signed an MOU to develop an energy-industry specific large language model.
Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
No API provider serves it, so its downloads are people running it themselves: 5 quantized builds and 6 Spaces.
Top Spaces running it: multimodalart/jev-decision-index ♥ 802
Weights only, an estimate: leave room for the KV cache. Method
About 2.7 GB at INT4, 5.4 GB at FP8, 11 GB at BF16: weights only, so leave room for the KV cache.
Estimate params × bytes per param × 1.2 · how we estimate
Weights, every dated article, and repo commits on one rail.
Teal is the weights (2026-09-24), amber triangles are 1 dated article (each links to it), grey ticks are repo commits.
5 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Everything else, folded: the full tables, every source.
| Benchmark | Value | Setting / notes | Reported by | Date |
|---|---|---|---|---|
| author-reported Language ACOS | 17.5 | — | model author | 2026-10-07 |
| author-reported Language ANLI | 43.1 | — | model author | 2026-10-07 |
| author-reported Language ContractNLI | 70.6 | — | model author | 2026-10-07 |
| author-reported Language FinEntity | 88.1 | — | model author | 2026-10-07 |
| author-reported Language NLI4CT | 58.1 | — | model author | 2026-10-07 |
| author-reported Language RAGTruth | 46.5 | — | model author | 2026-10-07 |
| author-reported Language VAST | 45.1 | — | model author | 2026-10-07 |
| author-reported Language iSarcasmEval | 37.9 | — | model author | 2026-10-07 |
| author-reported MMLU-Pro · Reason through hard problems | 42.5 | — | model author | 2026-10-07 |
| author-reported Retrieval Amazon ESCI | 32.8 | — | model author | 2026-10-07 |
| author-reported Retrieval CLINC150+OOS | 64.3 | — | model author | 2026-10-07 |
| author-reported Terminal-Bench · Write code with agents | 24.5 | — | model author | 2026-10-07 |
| author-reported Winogrande · Chat and follow instructions | 42.5 | — | model author | 2026-10-07 |
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
Listed in Decision models
| Pipeline tag | image-text-to-text |
|---|---|
| Library | transformers |
| Architectures | Qwen3_5ForConditionalGeneration |
| Model type | qwen3_5 |
| Dtype | bfloat16 |
| Params by dtype | BF16: 4,539,261,696 · F32: 3,840 |
| Size band | M |
| MoE | no |
| Context length | 262,144 (config.json max_position_embeddings) |
| License | CC BY-NC 4.0 (cc-by-nc-4.0) |
| Hub snapshot | 2026-10-25 |
vLLM: supported · Qwen3_5ForConditionalGeneration · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Qwen3_5ForConditionalGeneration · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed
vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.
Weights-only memory for jpt-4b: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 11 GB · Single GPU (<24 GB) |
|---|---|
| FP8 (1 byte/param) | 5.4 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 2.7 GB · Edge (<4 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
| How it was made | LoRA/adapter fine-tune |
|---|---|
| Lineage bucket | F - fine-tune of an open model (public/synthetic data) |
| Lineage evidence | board file inspection (adapter/head/weight files) + HF base_model field |
| Base model | Qwen/Qwen3.5-4B-Base |
| DI 0.2.1 score | 43.04 (independent (DI 0.2.1 board)) |
| ECE (board) | 0.0793 |
| Median latency | 139.1 ms (DI board, 1x RTX PRO 6000, single process) |
| Hosted price (input) | self-hosted/M |
| Source | https://huggingface.co/kirp/jpt-4b |
Base model: Qwen/Qwen3.5-4B
quantized: 5
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| prithivMLmods/jpt-4b-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | prithivMLmods | 3 | 1.4k |
| DreamBlooms/jpt-4b-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | DreamBlooms | 0 | 183 |
| ViniciosLugli/jpt-4b-nvfp4 | quantized | nvfp4 | 2.76B | M | cc-by-nc-4.0 | ViniciosLugli | 1 | 39 |
| RepublicOfKorokke/jpt-4b-oQ4e-fp16 | quantized | -fp16 | 4.54B | M | — | RepublicOfKorokke | 1 | 38 |
| RepublicOfKorokke/jpt-4b-oQ4e | quantized | — | 4.54B | M | — | RepublicOfKorokke | 1 | 37 |
| Date | Commit | SHA |
|---|---|---|
| 2026-09-27 | Model card: add JPT-35B-A3B to the family links, JevBench and Decision Index | 78312f8 |
| 2026-09-27 | Model card: Jev 1.13.0 official scores as baseline; Decision Index by area vs Jev | 395c19c |
| 2026-09-27 | Model card: fix badge anchors | 4560c14 |
| 2026-09-27 | Model card: new layout (charts, badges, Decision Index 0.2.1, versioned benchmarks with references) | 4d35e2b |
| 2026-09-27 | Decision Index 40.33 -> 39.54 (corrected run: 13,009 rows had been served by JPT-9B) | 05285d2 |
| 2026-09-25 | Add Decision Index 0.2 score | 277ae37 |
| 2026-09-24 | Card: v11 results, NeoHorse head-to-head (JevBench / Kev / OpenJev), image rows; GIFs re-recorded with v11 | 0ac3499 |
| 2026-09-24 | JPT-4B v11 weights (v10 + constrained-tradeoff documents): JevBench public 0.879, hard 0.784 | 13f4884 |
| 2026-09-24 | Demo GIFs: wrapped text, readable font, square-ish frames | 2792573 |
| 2026-09-24 | Model card with v10 results, JevBench v1.4 comparison, demo GIFs | 0d4ec2f |
| 2026-09-24 | JPT-4B v10 weights (merged Qwen3.5-4B LoRA, mix_train_env_v10) | d4e330b |
| 2026-09-24 | initial commit | b343d80 |
Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:44 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:18 UTC · Last verified 2026-10-25
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.