Jeeves

tom@tomrochette.com (Tom Rochette) — Tue, 06 Oct 2026 00:00:00 +0000

Jeeves is PostHog’s MIT-licensed research model, a 9B Jev-style decision model (a LoRA and pointer head on Qwen3.5-9B plus a small diffusion drafter) trained with SFT and CISPO to reason briefly before it answers typed questions.

Jeeves is the first entrant to attack the wave’s core weakness, that single-pass decision models trade accuracy for speed, and its answer, think for about 3 seconds before deciding, beats hosted Jev’s published numbers on its own tests while conceding the speed the contract was built for.

What it is
#

A research artifact from PostHog, the product-analytics company, acknowledging Kev as its inspiration: the same LoRA-plus-pointer-head architecture, plus a block-4 diffusion drafter, trained with supervised fine-tuning and CISPO so the model drafts a short reasoning chain and then reads out its answer. It supports noul, choice, and score questions in one request through a Jev-compatible local server, runs on CUDA in bf16 or fp8 and on Apple Silicon through MPS, and ships full training code, train/dev/test data, and Apache-2.0 weights on Hugging Face. Latency is the trade: about 0.3 s per request without thinking and a 3.3 s median with it on one H100 in fp8; on an M4 Mac the thinking runs at about 20 tokens per second.

Status
#

Days old and a research result, not a product. The repository was created 2026-09-29 and pushed 2026-10-01, with 411 stars and 21 forks as of 2026-10-06. The launch thread (2026-09-29) reached 242 points, and Hugging Face shows 216 downloads and 5 likes. Every benchmark in the README is self-run: on its held-out test split Jeeves scores 0.889 against Kev-9B’s published 0.822 and Jev’s published 0.857, and 0.935 against Jev’s 0.866 on JevBench’s 231 public items, while Jev keeps the transfer lead (0.800 against 0.746 on MMLU-Pro and buried state) and the same checkpoint without thinking drops to 0.804. JevBench’s own board does not list Jeeves, so the public-tier numbers remain author-run.

Strengths
#

Reasoning composes with the decision contract: thinking lifts the same checkpoint from 0.804 to 0.889 on its test split, the clearest published answer yet to whether the single-pass accuracy ceiling is architectural or just untrained.
The unknown-question discipline beats the original: 0.055 of unknowable questions answered at p 0.9 or higher against Jev’s published 0.090, with better calibration error than Jev on public items (0.037 against 0.049).
Everything needed to check it ships: training code, data, drafter, and a server, so the tables are re-runnable in a way most launch posts are not.
Third-party adoption arrived within a week: Ollaya packages jeeves:9b in its library.

Cautions
#

Every comparison is self-run against published numbers measured on different samples, the caveat this wave keeps failing to escape, and the benchmark’s sealed tiers have never seen Jeeves.
Thinking costs roughly ten times the hosted contract’s latency (3.3 s median on an H100), which defeats the millisecond use case the category exists for; the result is a dial, not a free lunch.
Apple silicon needs about 48 GB to avoid swapping with default caches.
The thread’s skeptics landed the same two punches the whole wave absorbs: one called the primary use case “posting on social media about Jev”, and others asked why not just fine-tune a small model, with no independent evaluation yet on either side.

Pricing
#

Free and open: MIT-licensed code and Apache-2.0 weights, no hosted service and no paid tier. The cost is a CUDA GPU (or a large Mac) and the patience to wait out the thinking budget.

Compared to
#

Kev: the acknowledged inspiration; Kev ships the locked-test eval discipline and the wider checkpoint family, Jeeves ships the reasoning result on top of the same architecture.
Jev: the closed original still wins transfer, latency, and independent scrutiny; Jeeves wins its own held-out test and the public JevBench items it ran.
Jeff: the other second-generation student line; Jeff’s v1.3 adapters decide when a big model should think, Jeeves makes the small model think itself, and both keep the calibration-first framing.

Bottom line
#

Recommended for researchers of the decision contract who want to test whether brief reasoning generalizes, and for PostHog-scale teams with GPUs to spare. Not as a production dependency today: three weeks old, one company behind it, and no number a second party has confirmed. The disagreeable claim I will defend: if brief reasoning generalizes across the wave, the single-pass architecture stops being the point and the contract becomes any calibrated classifier with a thinking budget, which is bad news for everyone selling speed as the moat.

Changes
#

2026-10-06 - Created from the entrant scan after the 2026-09-29 Show HN thread cleared the bar (242 points, PostHog standing, full training artifacts shipped).

References
#

https://github.com/PostHog/jeeves - repository: MIT, created 2026-09-29, 411 stars, 21 forks, pushed 2026-10-01 (GitHub API, as of 2026-10-06)
https://raw.githubusercontent.com/PostHog/jeeves/master/README.md - the benchmark tables, the LoRA-plus-pointer-head-plus-drafter architecture, CISPO training, latency table, and the Kev acknowledgement
https://news.ycombinator.com/item?id=49891290 - the launch thread (242 points as of 2026-10-06, 2026-09-29), including the copycat and social-media skepticism
https://huggingface.co/PostHog/jeeves - weights: Apache-2.0, created 2026-09-29, 216 downloads and 5 likes as of 2026-10-06 (Hugging Face API)
https://ollaya.dev/ - third-party adoption: jeeves:9b in the Ollaya library and its accuracy table

reasoning on tomrochette.com

Jeeves

What it is
#

Status
#

Strengths
#

Cautions
#

Pricing
#

Compared to
#

Bottom line
#

Changes
#

See also
#

References
#

reasoning on tomrochette.com

Jeeves

What it is #

Status #

Strengths #

Cautions #

Pricing #

Compared to #

Bottom line #

Changes #

See also #

References #

What it is
#

Status
#

Strengths
#

Cautions
#

Pricing
#

Compared to
#

Bottom line
#

Changes
#

See also
#

References
#