╭────────────────────────────────────────────────────────────────────────╮
│                                                                        │
│  R I G M A R K   //   AGENT WORKLOAD RECEIPT                           │
│  BENCHMARKS LOCAL AI HOW CODING AGENTS ACTUALLY USE IT                 │
│  ▲  9/15 BASIC OUTPUT GATES PASSED — DO NOT HEADLINE                   │
│                                                                        │
├─ MODEL ────────────────────────────────────────────────────────────────┤
│                                                                        │
│  GLM-5.3-Flash-EXL3                                                    │
│                                                                        │
├─ SINGLE STREAM ────────────────────────────────────────────────────────┤
│                                                                        │
│  WORKLOAD         tok/s       Range tok/s   Last (s)    Checks         │
│  PROSE             62.5         56.5–66.9      108.0     ✗ 4/5         │
│  CODE              67.8         65.2–72.7      121.3     ✗ 0/5         │
│  STRUCTURED*      102.6       101.2–103.4       10.1     ✓ 5/5         │
│                                                                        │
│  * Predictable JSON ceiling; not general agent performance.            │
│  Decode medians are estimates; rates include streamed reasoning.       │
│                                                                        │
├─ PREFILL ──────────────────────────────────────────────────────────────┤
│                                                                        │
│  DEPTH       Cold tok/s      Cold TTFT (s)    Replay TTFT (s)          │
│  8K               2,028               4.04               0.04          │
│  32K              2,065              15.87               0.05          │
│  64K              2,041              32.12               0.06          │
│  Medians; replay is an immediate repeat, not a proven hit.             │
│                                                                        │
├─ CAPPED CONCURRENT GENERATION ─────────────────────────────────────────┤
│                                                                        │
│  Aggregate tok/s: C1 56.9 | C2 83.4 | C4 130.1                         │
│  C4: 0/12 normal stops; 0/12 with visible output.                      │
│  Capped throughput includes reasoning; not completed agent tasks.      │
│                                                                        │
├─ APPLIANCE ────────────────────────────────────────────────────────────┤
│                                                                        │
│  Hardware: 2x NVIDIA DGX Spark (GB10, 128 GB unified memory each),     │
│    ascent-05 + ascent-06                                               │
│  Topology: TP2, one rank per Spark, RoCE over ConnectX-7               │
│  Checkpoint: Mia-AiLab/GLM-5.3-Flash-EXL3-4bpw-TensorFold @ HF snapshot│
│  Quantisation: EXL3 4 bpw weights, BF16 activations                    │
│  KV cache: engine default for this lane                                │
│  Engine: TensorFold (image tensorfold-glm53:v1.10-fleet-64k)           │
│                                                                        │
├─ SETTINGS ─────────────────────────────────────────────────────────────┤
│                                                                        │
│  SUITE: DEFAULT SETTINGS                                               │
│  REQUEST: {}                                                           │
│  Temperature 0.0 | top_p 1.0 | seed 20260905 | protocol 1.3.0          │
│  Decode: 5 runs, 8192-token cap | Prefill: 8K/32K/64K, 3 pairs         │
│  Concurrency: C1/C2/C4, 3 rounds, 256-token cap; workload code         │
│  Cache isolation: random-salt-per-pair                                 │
│  Comparison ID: 2026-10-two-sparks-four-models-v1                      │
│                                                                        │
├─ RECEIPT ──────────────────────────────────────────────────────────────┤
│                                                                        │
│  SOURCE     git:946394fa53db  •  clean                                 │
│  RIGMARK RELEASE: ahead of v0.0.4 (unreleased checkout)                │
│  JSON sha256:58b3aacb21e25a9d…                                         │
│  SHARE THE CARD • LINK THE JSON RECEIPT • #RIGMARK                     │
│  github.com/alexellis/rigmark                                          │
│                                                                        │
╰────────────────────────────────────────────────────────────────────────╯