Skip to content
brianletort.aiD. Brian Letort, Ph.D.

Model layer

The Model Pulse

October 24, 2026No. 27

For architects tracking model capability shifts.

Complete issue archive.

Model lineage, architecture, benchmarks, vendor signals, and tree deltas in one software-layer read.

Subscribe via RSS

Published issues

27 issues · newest first

  1. No.27

    Week 43 of 2026 · October 24, 2026

    A fourth lab reached the $2 / $10 band on an API; per completed task it costs about three times the band's cheapest member and about a third of Opus

  2. No.26

    Week 42 of 2026 · October 17, 2026

    Three frontier-scale open checkpoints in 72 hours, and the model card is no longer the number to plan on

  3. No.25

    Week 41 of 2026 · October 10, 2026

    Every price that fell this week fell for a segment; key your router on prompt length and call type, not list price

  4. No.24

    Week 40 of 2026 · October 3, 2026

    Same list price, three different bills

  5. No.23

    Week 39 of 2026 · September 26, 2026

    Two priced models shipped, and the one that did not is the one OpenAI paused

  6. No.22

    Week 38 of 2026 · September 19, 2026

    Voice models stopped waiting for tools, and the architecture decision moved to concurrency, metering, and trust boundaries

  7. No.21

    Week 37 of 2026 · September 12, 2026

    The serving layer changed underneath models already deployed, and this week's new checkpoints lead on cost and footprint rather than capability

  8. No.20

    Week 36 of 2026 · September 5, 2026

    Model governance now starts at the deployable package: entitlement, retention, runtime version, and price window

  9. No.19

    Week 35 of 2026 · August 29, 2026

    Kodam's practitioner harness result and seat metering, not Hot Chips silicon — the week's model procurement read

  10. No.18

    Week 34 of 2026 · August 22, 2026

    OpenAI cut Sol by more than a fifth for three months while an unknown lab shipped an MIT frontier — the same week the industry conceded that vendor-reported is what the top of the board looks like now

  11. No.17

    Week 33 of 2026 · August 15, 2026

    Six releases in five days, and the model you can benchmark stopped being the model you can run

  12. No.16

    Week 32 of 2026 · August 8, 2026

    Two new frontier models shipped and the more important releases were the measuring instruments — the same weights now score differently depending on who serves them

  13. No.15

    Week 31 of 2026 · August 1, 2026

    Open weights finally became deployable artifacts — and the closed stack answered with post-training and rate cards, not new flagships

  14. No.14

    Week 30 of 2026 · July 25, 2026

    Opus-class intelligence just moved into the mid-tier price band, while Google reset the fleet-economics floor

  15. No.13

    Week 29 of 2026 · July 18, 2026

    Kimi K3 topped an LMArena flagship board, priced itself onto the closed vendors' rate card — and shipped without weights. Meanwhile the largest US-origin open-weights model actually landed, and OpenAI showed what it keeps for itself.

  16. No.12

    Week 28 of 2026 · July 11, 2026

    Four frontier releases in four days — and the week's real lesson came from the community: Hy3 went from Apache-2.0 drop to local deployment in 30 hours, while Databricks proved the harness now matters more than the model.

  17. No.11

    Week 27 of 2026 · July 4, 2026

    Washington entered the release pipeline: GPT-5.6 launched gated to ~20 approved partners, Fable 5 came back, and Sonnet 5 split sticker price from cost-per-task.

  18. No.10

    Week 26 of 2026 · June 27, 2026

    No new model reset the board; the model-layer story moved to availability, cost, and serving economics.

  19. No.09

    Week 25 of 2026 · June 20, 2026

    Open weights took the lead while the closed frontier stalled — and the benchmark itself was rewritten.

  20. No.08

    Week 24 of 2026 · June 13, 2026

    A new closed frontier shipped and was switched off in the same week; the durable open progress was a diffusion efficiency path.

  21. No.07

    Week 23 of 2026 · June 6, 2026

    No new closed frontier shipped; the open/local agent substrate widened underneath it.

  22. No.06

    Week 22 of 2026 · May 30, 2026

    A quiet release week, a loud leaderboard: Anthropic retook the frontier with Opus 4.8.

  23. No.05

    Week 21 of 2026 · May 23, 2026

    Six new tree rows. The loudest model-layer week of Q2 — Gemini 3.5 + Omni, Qwen3.7-Max in the global top 5, three open-weights drops that reset what 'open' means.

  24. No.04

    Week 20 of 2026 · May 17, 2026

    Frontier text took a breath; the specialist canopy widened — video, on-device multimodal, open world-models — and agent-runtime monetization became the story.

  25. No.03

    Week 19 of 2026 · May 8, 2026

    Frontier text went quiet. Voice crossed into chain-of-thought. The week's biggest capability move was a compute deal.

  26. No.02

    Week 18 of 2026 · May 1, 2026

    Six new tree rows, two distribution earthquakes, and the open-weights ceiling closing to 6 points behind GPT-5.5.

  27. No.01

    April 2026 Recap · April 25, 2026

    April rewrote the floor: open weights crossed the closed frontier, and the safety-gated frontier became a procurement criterion.

Methodology

What survives into every issue

  • Every issue starts with the model-tree delta since the prior issue.
  • Frontier and open-weight releases retain tier, architecture, evidence, and source context.
  • Benchmark moves cite public leaderboards and model cards as of publish date.
  • All sources are public. Independent analysis only.