Model economics / detail
deepseek-v3D
Dispatch tier: Inexpensive (complexity class this model is routed for)
Two runs, zero tokens, dead in 8 seconds each — this isn't a model, it's a misconfigured endpoint wearing a name tag. 0% success is unforgiving, even if the wiring, not the weights, deserves the blame.
The measured record
What the ledger shows
Fleet-ledger aggregates for this model only — run counts, outcomes, spend, and token appetite. Missing telemetry reads “Not observed”, never zero.
Agent runs
2
Success rate
0%
Metered spend
Not observed
cost-blind: no run was metered
Cost / metered run
Not observed
Tokens in
0
across 2 token-measured runs
Tokens out
0
Tokens in / run
0
Tokens out / run
0
Outcomes
Status breakdown
Every run ends in exactly one status. The bar is the whole record, to scale.
- failure: 2 of 2 runs (100.0% of total)
Run duration
How long the runs take
Wall-clock distribution across measured runs, from the fastest exit to the longest grind.
Min
8s
p25
8s
Median
8s
p75
8s
p90
8s
p99
8s
Max
8s
Mean
8s
Duration measured on 2 of 2 runs.
Commentary
Idiosyncrasies
What stands out in this model's numbers — shape, appetite, and failure habits.
- Two runs, both dead at 8 seconds, zero tokens in or out — the endpoint never answered.
- 0% success with a token counter that reads exactly zero is a connectivity signature, not a capability one.
Commentary
Lessons learned
Practical routing and operations takeaways, grounded in the same record.
- Nothing to learn yet beyond: fix the router (or the endpoint config) — the model never got a word in.