Speed sweep
qwen3-6-27b-q4-k-m-rtx-4090-24gb-llama-cpp-tp1-jarmcutc-sweep
qwen3-6-27b-q4-k-m-rtx-4090-24gb-llama-cpp-tp1-jarmcutc-sweepRecord
- Measured
- 2026-04-30T19:36:40.754Z
- Points
- 1
Measured speed
| Concurrency | Context | Prefill | Decode | TTFT ms | Status |
|---|---|---|---|---|---|
| 1 | 90,200 | — | 46.6 | 409 | observed |
Remaining fields
Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.
- id
- qwen3-6-27b-q4-k-m-rtx-4090-24gb-llama-cpp-tp1-jarmcutc-sweep
- measured at
- 2026-04-30T19:36:40.754Z
- recipe id
- qwen3-6-27b-q4-k-m-rtx-4090-24gb-llama-cpp-tp1-jarmcutc
- schema version
- local-ai-registry/v1
metrics
- concurrency
- 1
- inference engine version
- b8818-b572d1ecd
- latest point at
- 2026-04-30T19:36:40.754Z
- max context tokens
- 90,200
- peak generation tps
- 46.64
- point count
- 1
rows
| concurrency | context tokens | decode tok s | decode tok s per stream | output tokens | peak vram gb | prefill tok s | samples |
|---|---|---|---|---|---|---|---|
| 1 | 90,200 | 46.64 | 46.64 | 512 | 18.3 | Unknown | 1 |
Provenance & metadata (1)
source
- kind
- leaderboard
- repository
- www.localmaxxing.com ↗