350-401 ENCOR Sample Questions and the Passing Score Explained
The passing score, without the folklore
Here is the honest answer first: Cisco does not publish a fixed passing percentage for the 350-401 ENCOR exam. Cisco reports your result on a scaled score from 300 to 1,000, and the cut score on that scale is set by Cisco, can differ between exam forms, and is not published. When you finish, you receive a pass or fail decision plus a score report showing your percentage performance in each section of the blueprint. The exact figures repeated in forums and older blog posts are anecdotes from individual score reports, not a Cisco standard — and nobody can aim at a number the vendor will not confirm, so prepare to be strong across the whole blueprint instead.
The format, as Cisco currently describes it, is approximately 100 questions in 120 minutes — Cisco does not publish an exact question count — which works out at roughly 72 seconds per question, with a heavy share of exhibits that must be read before the options make sense. The five questions below are original to this guide: a symptom, several plausible responses, and one answer Cisco's way of thinking rewards.
Five sample questions, worked properly
Sample question 1. OSPF is enabled on both interfaces of a point-to-point link; both are in area 0 and the same subnet, and authentication is not configured — yet the adjacency never reaches the FULL state. Which mismatch is a classic cause?
- A. The routers use different OSPF process IDs
- B. The hello and dead intervals on the link do not match
- C. The routers have different router IDs
- D. One router has more OSPF-enabled interfaces than the other
Answer: B. Hello and dead intervals are carried in the hello packets and must agree, or the routers decline to become neighbours — this sits on the standard adjacency checklist alongside area ID, subnet and mask, MTU, and stub flags. The strongest distractor is A: the OSPF process ID is locally significant only, so different process IDs on different routers are perfectly normal. Different router IDs (C) are required, not a fault; a duplicate router ID is the problem. Interface count (D) is irrelevant to a single adjacency.
Sample question 2. A campus LAN reconverges every time a distribution uplink flaps, and traffic floods while it does. Investigation shows an old access switch has become the spanning-tree root bridge because nobody ever configured priorities. Which change addresses the root cause?
- A. Enable PortFast on the distribution uplink ports
- B. Deliberately set a core or distribution switch as the root bridge with a low priority, and use PortFast with BPDU guard on access edge ports only
- C. Disable spanning tree on the access switches
- D. Increase the spanning-tree hello time on every switch
Answer: B. Root placement is a design decision: the root belongs at the centre of the topology, chosen deliberately, not left to whichever switch wins an election on default values. The strongest distractor is A, and it is a trap — PortFast on a switch-to-switch uplink skips the listening and learning states, inviting temporary loops and topology-change storms, the very symptoms described. PortFast belongs on edge ports facing end devices, protected by BPDU guard so a rogue switch plugged into one is shut down instead of joining the topology. Disabling spanning tree (C) removes the safety net entirely.
Sample question 3. An engineer is reading data returned by a network device's API. Which snippet correctly represents a JSON object holding a list of interface names under the key "interfaces"?
- A. "interfaces": "GigabitEthernet0/1, GigabitEthernet0/2"
- B. "interfaces": ["GigabitEthernet0/1", "GigabitEthernet0/2"]
- C. "interfaces": {GigabitEthernet0/1, GigabitEthernet0/2}
- D. interfaces = (GigabitEthernet0/1; GigabitEthernet0/2)
Answer: B. In JSON, square brackets denote an array — an ordered list of values, each a quoted string here. Option A is a single string containing a comma; a script reading it receives one value, not two. Option C uses braces, which denote an object — and its unquoted names are invalid JSON anyway. Option D is not JSON syntax at all. ENCOR's automation section expects exactly this level of data-format literacy: recognising structures at a glance, in JSON and in its YAML equivalents.
Sample question 4. Voice calls across a WAN link become choppy whenever a bulk file transfer saturates the link. The circuit cannot be upgraded this quarter. Which QoS design protects the voice traffic?
- A. Order more bandwidth for the circuit
- B. Classify and mark the voice traffic, then serve it from a strict-priority low-latency queue while the bulk traffic shares what remains
- C. Enable first-in, first-out queuing on the WAN interface
- D. Apply the same DSCP marking to all traffic so nothing is treated unfairly
Answer: B. Congestion management is the mechanism being tested. Voice is delay- and jitter-sensitive, so it is classified and marked, then served from a low-latency queue — with policing so it cannot starve everything else — while bulk traffic absorbs the delay instead. Option A is the strongest distractor because it sounds like common sense, but the scenario rules it out, and Cisco's credited answers favour the QoS mechanism that protects voice during congestion at any circuit size. FIFO (C) is the absence of a policy: voice waits in line behind the file transfer. Marking everything identically (D) destroys the differentiation QoS exists to create.
Sample question 5. A router learns the prefix 10.10.0.0/16 from two sources. EIGRP reports it with a metric that looks poor; OSPF reports it with a metric that looks excellent. Which route does the router install?
- A. The OSPF route, because its metric is better
- B. The EIGRP route, because administrative distance — not metric — decides between different routing sources
- C. Both routes, load-balanced equally
- D. Whichever route was learned first
Answer: B. Metrics are only ever compared between routes learned from the same protocol. When the same prefix arrives from different sources, the router first compares administrative distance — its ranking of how trustworthy each source is. EIGRP's internal administrative distance of 90 beats OSPF's 110, so the EIGRP route installs despite the uglier metric; only if two sources tie on distance does the metric comparison begin. Option A is the classic trap: it treats the metric as a universal score. Arrival order (D) plays no part in the decision.
How to use questions like these
Two habits separate ENCOR candidates who pass from those who nearly pass. The first is exhibit discipline. At roughly 72 seconds per question, you cannot afford to reread a topology three times: read the stem first, then mine the exhibit for exactly the values the question needs, then commit to an answer — reading every exhibit line by line is how candidates run out of clock. The second habit is diagnostic honesty after practice. Your real score report breaks performance down by blueprint section, so make your practice do the same — tag every miss to its section across architecture, virtualisation, infrastructure, network assurance, security, and automation, then repair the weakest section instead of simply doing more questions.
One warning about sources: anyone selling you the actual exam questions is selling stolen material, and using it risks Cisco revoking the very credential you are working toward — whether 350-401 practice tests are accurate and safe explains how to tell the difference. Original questions in the exam's style train the same judgement without that risk — and timed, domain-scored free practice tests on Study4Pass let you rehearse the 72-seconds-per-question pace. Two practicalities complete the picture: how much the 350-401 exam costs and whether ENCOR includes live labs.