Tiiny Pocket AI box tested: gpt-oss-120b hits 16 tok/s and is the only model to pass a 10-test coding benchmark
IronDiligent6698 · reddit · 2026-11-05
- Hardware: Tiiny Pocket is 5.6×3.1×0.9 inches, 300g, with 32GB SoC memory plus 48GB dedicated NPU memory (80GB total, not unified), bought at an early-backer price around $1K.
- Speed (4K input, prompt caching off, median of 5 runs, Qwen3.5-9B-Turbo INT4): 1041 tok/s prompt processing, 75 tok/s generation, 4.05s TTFT; Qwen3.6-35B-A3B-Turbo manages 64 tok/s; gpt-oss-120b (mixed INT4/INT8, 67.3GB) only 16 tok/s, taking 30+ seconds per request.
- Coding test: identical Python inventory/order management task graded by 10 automated tests. The 9B never produced a gradable answer; the 35B MoE scored 9/10, missing one edge case on retrying a sold-out order; gpt-oss-120b was the only model to pass 10/10, in 229 seconds — nearly the same wall time as the 35B's 235 seconds.
- Takeaway: small models are fast but unreliable; on-device 120B is slow yet the only one that completes a real coding task.
More from Embodied
- Beff Jezos shares IRL photos of Tesla Cybertruck at NASA Florida event — beffjezos · 2026-11-05
- Waymo to end Uber exclusivity in Austin and Atlanta, launch own app in January 2028 — emmanuelvivier · 2026-11-05
- Tired of screens and headsets, this maker built a wearable computer hat — aaraalto · 2026-11-05
- Evopien: local-first conversational AI prototype with vision runs on Jetson AGX Thor — SoftKill21 · 2026-11-05
- DJI Matrice 5T packs 1K thermal, LiDAR and 59-min flight into one enterprise drone — davidpattersonx · 2026-11-05
- Robotics strategy: be the last mover and capture monopoly profits — StewartalsopIII · 2026-11-05