Up to date
kirp CC BY-NC 4.0 out 30 days ago
9.41B 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 1.1k · #591 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: 3 quantized builds and 7 Spaces.
Top Spaces running it: multimodalart/jev-decision-index ♥ 802 · kirp/jpt-9b-typed-decisions ♥ 1
Weights only, an estimate: leave room for the KV cache. Method
About 5.6 GB at INT4, 11 GB at FP8, 23 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 (2026-09-25), 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 ACOS | 20.3 | Language | model author | 2026-10-07 |
| author-reported ANLI | 50 | Language | model author | 2026-10-07 |
| author-reported Amazon ESCI | 40.6 | Retrieval | model author | 2026-10-07 |
| author-reported Arts | 28.6 | — | model author | 2026-10-07 |
| author-reported BBH · Reason through hard problems | 55.6 | Knowledge | model author | 2026-10-07 |
| author-reported BPoMP | 72.4 | Arts | model author | 2026-10-07 |
| author-reported BRIGHT · Rerank search results | 39.9 | Retrieval | model author | 2026-10-07 |
| author-reported Banking77 · Classify and make typed decisions | 65 | Retrieval | model author | 2026-10-07 |
| author-reported CLINC150+OOS | 35.8 | Retrieval | model author | 2026-10-07 |
| author-reported CLadder | 33.8 | Knowledge | model author | 2026-10-07 |
| author-reported CRUXEval | 29.6 | Knowledge | model author | 2026-10-07 |
| author-reported ChessBench | 7.3 | Games | model author | 2026-10-07 |
| author-reported ContractNLI | 74.2 | Language | model author | 2026-10-07 |
| author-reported Decision Index (DI) · Classify and make typed decisions | 46.89 | — | model author | 2026-10-07 |
| author-reported FinEntity | 90 | Language | model author | 2026-10-07 |
| author-reported GPQA Diamond · Reason through hard problems | 21.8 | Knowledge | model author | 2026-10-07 |
| author-reported GSM8K · Reason through hard problems | 69.2 | Knowledge | model author | 2026-10-07 |
| author-reported Habermas Machine | 13.2 | Arts | model author | 2026-10-07 |
| author-reported HellaSwag · Chat and follow instructions | 81.2 | Language | model author | 2026-10-07 |
| author-reported HoVer | 32.7 | Retrieval | model author | 2026-10-07 |
| author-reported Humanity's Last Exam · Reason through hard problems | 0 | Knowledge | model author | 2026-10-07 |
| author-reported Humicroedit | 19.5 | Arts | model author | 2026-10-07 |
| author-reported Knowledge | 31.7 | — | model author | 2026-10-07 |
| author-reported Language | 56.7 | — | model author | 2026-10-07 |
| author-reported MMLU-Pro · Reason through hard problems | 49.1 | Knowledge | model author | 2026-10-07 |
| Benchmark | Value | Setting / notes | Reported by | Date |
|---|---|---|---|---|
| author-reported MuSR | 28.1 | Knowledge | model author | 2026-10-07 |
| author-reported NLI4CT | 59.1 | Language | model author | 2026-10-07 |
| author-reported New Yorker | 52.6 | Arts | model author | 2026-10-07 |
| author-reported POP909-CL | 3.5 | Arts | model author | 2026-10-07 |
| author-reported RAGTruth | 50.6 | Language | model author | 2026-10-07 |
| author-reported Retrieval | 44.6 | — | model author | 2026-10-07 |
| author-reported SATA-Bench | 22.6 | Knowledge | model author | 2026-10-07 |
| author-reported Terminal-Bench · Write code with agents | 27.4 | Arts | model author | 2026-10-07 |
| author-reported Tools | 67 | — | model author | 2026-10-07 |
| author-reported VAST | 46.6 | Language | model author | 2026-10-07 |
| author-reported Winogrande · Chat and follow instructions | 49.2 | Language | model author | 2026-10-07 |
| author-reported cfcolor | 11.4 | Arts | model author | 2026-10-07 |
| author-reported iSarcasmEval | 43.9 | Language | 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: 9,409,809,904 · F32: 3,840 |
| Size band | L |
| 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-9b: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 23 GB · Single GPU (<24 GB) |
|---|---|
| FP8 (1 byte/param) | 11 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 5.6 GB · Single GPU (<24 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/Qwen3.5-9B-Base |
| DI 0.2.1 score | 46.89 (independent (DI 0.2.1 board)) |
| ECE (board) | 0.0773 |
| Median latency | 148.6 ms (DI board, 1x RTX PRO 6000, single process) |
| Hosted price (input) | self-hosted/M |
| Source | https://huggingface.co/kirp/jpt-9b |
Base model: Qwen/Qwen3.5-9B
quantized: 3finetune: 2
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| prithivMLmods/jpt-9b-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | prithivMLmods | 3 | 1.3k |
| RepublicOfKorokke/jpt-9b-oQ4e | quantized | — | 9.41B | L | — | RepublicOfKorokke | 0 | 113 |
| RepublicOfKorokke/jpt-9b-oQ4e-fp16 | quantized | -fp16 | 9.41B | L | — | RepublicOfKorokke | 0 | 72 |
| sergolsub/mlx-jpt-9B | finetune | mlx | 9.41B | L | cc-by-nc-4.0 | sergolsub | 0 | 9 |
| sergolsub/mlx-jpt_9B | finetune | mlx | — | — | cc-by-nc-4.0 | sergolsub | 0 | 0 |
| Date | Commit | SHA |
|---|---|---|
| 2026-10-04 | Model card: AG News / Emotion / SST-5 train splits are in the training mix (not out-of-distribution) | 9fda2c7 |
| 2026-09-27 | Model card: JPT-9B JevBench public 0.853 measured with the JevBench CLI | b447cc7 |
| 2026-09-27 | Model card: add JPT-35B-A3B to the family links, JevBench and Decision Index | dd416ab |
| 2026-09-27 | Model card: Jev 1.13.0 official scores as baseline; Decision Index by area vs Jev | e389a7c |
| 2026-09-27 | Model card: fix badge anchors | 2b96880 |
| 2026-09-27 | Model card: new layout (charts, badges, Decision Index 0.2.1, versioned benchmarks with references) | a27d740 |
| 2026-09-25 | Add Decision Index 0.2 score | 7114b0c |
| 2026-09-25 | Add model card | d5e20b2 |
| 2026-09-25 | JPT-9B: mix_train_env_v11, LoRA r16 merged, chat prompt, T=1.087 | 2ad078e |
| 2026-09-25 | initial commit | fcee91c |
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 15:19 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.