Jebadiah 9B v2.1 has 9.197 billion total parameters. The selected Q8_0 GGUF file is 9.786 GB.
Data snapshot: . 1 published test in this catalog.
Model file and details
Provider
Frontier Infra
Total parameters
9,197,093,888
Exact file size
9,786,060,416 bytes
Quantization / precision
Q8_0
Weight repository
frontier-infra/jebadiah-9b-v2-1-GGUF
Original model repository
frontier-infra/jebadiah-9b-v2-1
Weight revision
11cd88b233b1227c2c49348f9994df776e51a2f3
Model revision
3ec34cf47c5bb9cce7ea32e59da17e9281341a08
Runtime notes
Confirm model architecture and quantization support in your runner.
Recorded release date
2026-10-10
Date evidence
Frontier Infra published the Apache-2.0 Jebadiah 9B v2.1 source checkpoint and official same-publisher GGUF set on 10 October 2026. Jebadiah is a text-only typed-decision model that returns probabilities over supplied labels; it is not a chat model. The selected 9,786,060,416-byte Q8_0 file is the highest-precision published GGUF and was checked against the merged bf16 weights on the first 260 presentations of the pinned A3 calibration split.
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.
Jebadiah 9B v2.1 Q8 · bf16 label agreement
98.8 / 100
Publisher-run top-label agreement for the selected Q8_0 GGUF against the merged bf16 weights on a fixed 260-question calibration sample. This measures quantization fidelity, not general accuracy or new-domain calibration.
Benchmark efficiency: 10.096 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 comparison of the exact selected Q8_0 GGUF against the merged bf16 weights on the first 260 question presentations from the pinned A3 calibration split. CUDA llama-server used llama.cpp commit aa94f2086147cd4bfbcff818ce998dffce5fcb39 on an RTX PRO 6000 with a 4,096-token server context; prompt token IDs matched the Python renderer. Q8_0 retained 257 of 260 top labels overall, including 155 of 156 choice/noul labels and 102 of 104 score labels, with maximum absolute probability difference 0.0459.
Publisher measurement, not independent validation. The sample is a calibration split and differs from the previous release's Jevals/Nimble sample, so cross-release agreement counts are not directly comparable. Agreement is not task accuracy or proof of calibration on a new domain; three top labels still flipped. Q5_K_M and Q4_K_M use less storage but changed 8 and 14 of 260 labels, respectively. The publisher's 47.20 Decision Index 0.3 public-only score and RTX PRO 6000 latency are full-weight measurements, not results for this GGUF, and the index run is not an accepted full leaderboard result. The selected file excludes runtime, KV cache, context and output headroom; file size alone does not establish minimum RAM, VRAM, throughput or stability. The exact build was checked with llama.cpp/CUDA, while Ollama and LM Studio were not checked for v2.1. Text-only training; no vision projector is shipped. 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 9.786 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 9.79 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 9.786 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.