Cloud baseline
Nemotron 3.5 Lightning — 0.0000 on Forge 1.0
nvidia/nemotron-3.5-lightning · scorer forge-1.0
Forge 1.0 was frozen on 9 October 2026 and accepts no new runs; Forge 2.0 is its successor. The original results contained scorer faults. Corrections made from saved evidence on 15 October are dated on the affected run pages, with both original and corrected scores and the reason for the change. They reuse the original builds and observations; no benchmark or model was rerun, and the original inference charges remain. Uncorrected results retain known limitations: the top five original scores, 0.945 to 0.977, were too close to reliably rank the strongest models; two hourly-job tests missed Jira reads performed by a queue worker; and the old linter accepted an entity index with two range attributes that Forge refuses. Those remaining faults are not silently treated as fixed. Consult each run's correction note when comparing historical results. Forge 2.0 has separate results and a different task.
- Overall
- 0.0000
- Benchmark
- Forge 1.0
- Model build
- 5m 46s
Recorded inference cost
$0.29
From the saved OpenRouter generation billing summary (151 recorded requests). Whether auxiliary model calls were included was not retained. Scoring reuses the saved build; this amount describes the original inference spend.
- Billing recovered 2026-10-18 from all 151 of this build's own generation ids with OpenRouter's generation records (151 billed, native token reconciliation passed). The earlier saved summary ($0.2887578, incomplete) was replaced; the benchmark score and run are unchanged.
Token usage
Tokens: 8.3M in · 53.4k out · 7.1M cached (exact, OpenRouter records)
8,344,658 tokens over 151 model calls: every call of the run, the model's own and the harness's helper calls, counted from OpenRouter records. Reasoning tokens as the provider reports them: 0 (not added to the total).
nvidia/nemotron-3.5-lightning-20260807
In 8.3M · Out 53.4k · Cached 7.1M · Reasoning 0 · 151 calls
Tier breakdown
Scoring detail
How this score was built
Per-test scorer output, as posted by the app. Expand a section to inspect its tests.
Correctness combines the lowest score for each test across the scoring fixtures; economy measurements use their mean. The combined result can therefore be lower than every individual fixture’s final score. Critical defects reduce the weighted result before any score cap applies.
Earned credit before admission
(0.88 × 0.0000 core + 0.12 × 0.00 gate × 0.00 excellence) × 0.3600 critical = 0.0000
Critical defects compound a multiplier on the whole score (pre-severity 0.0000):
l_deployable: no additional penalty (vacuous:precondition: >= 1 module declared); deploy blocked — lint errors.
l_bundles_load: no additional penalty (vacuous:precondition: >= 1 function declared); crash — the app does not run.
t_no_double_count: no additional penalty (vacuous:precondition: >= 1 ledger row); wrong numbers — a change counted twice.
b_no_permission_leak: no additional penalty (vacuous:precondition: >= 1 person-facing change list returned or rendered); data leak — hidden issues shown to a person.
b_comment_exactly_once: no additional penalty (vacuous:precondition: >= 1 comment POST); duplicate side effect on a customer's Jira.
The core (88% of the total) is the weighted mean of the nine measured tiers. The last 12% is the excellence slice: it unlocks in proportion to the perfection conditions below (0 of 18 met here), then pays out at the excellence tier's own measured mean. Core 0.0000 + excellence 0.0000.
Earned score 0.000 · Cap 0.499 · Final score 0.000
The final score is the lower of the earned score and the cap. Meeting a cap's conditions adds no points.
- deployable: maximum 0.499
l_deployablel_bundles_load - working ledger: maximum 0.699
u_widget_loadst_event_rowsr_backfill_completes_storage_scopes_entity_index_used - current platform, complete surfaces: maximum 0.799
k_dashboard_widgetk_widget_edit_bridgek_rovo_skillu_widget_edit_configu_ledger_tablea_action_resultv_theme_tokensv_dark_mode - production robustness: maximum 0.799
t_no_double_countt_out_of_ordert_retry_after_honouredr_heal_droppedr_rate_limitr_paginationb_no_permission_leakb_comment_exactly_onceb_realtime_payload_cleanu_llm_explainu_widget_livev_csp_cleanv_console_cleanu_ledger_sorts_index_orderb_hidden_counta_action_permissionsl_scopesk_llm_model_current
Forge's own linter reports no errors (warnings cost points), every manifest function bundles and loads, packages resolve at the kit's pins, the manifest rules the linter misses, and only the scopes the app's calls need.
l deployable
0.00vacuous — precondition unmet: >= 1 module declared
l bundles load
0.00vacuous — precondition unmet: >= 1 function declared
l lint warnings
0.00vacuous — precondition unmet: >= 1 function loads
l real packages
0.00vacuous — precondition unmet: >= 1 bundle imports @forge/*
l manifest rules
0.00vacuous — precondition unmet: >= 1 resource and >= 1 function
l scopes
0.00vacuous — precondition unmet: >= 1 product or KVS call observed
Current modules and APIs: dashboards:widget with its edit API, the rovo:skill and rovo:mcp wiring, a Forge LLM model the platform lists as active, no /rest/api/3/search, no @forge/api storage, a nodejs22.x or nodejs24.x runtime, the consumer shape, the declared KVS entity, and deploy readiness: a manifest and permissions real Forge would accept and no code predicted to fail there.
k dashboard widget
0.00required app surface absent: dashboards:widget module
k widget edit bridge
0.00vacuous — precondition unmet: edit surface rendered
k rovo skill
0.00required app surface absent: rovo:skill module
k current apis
0.00vacuous — precondition unmet: >= 1 product call observed
k consumer shape
0.00vacuous — precondition unmet: a consumer exists
k entity declared
0.00required app surface absent: app.storage.entities
k rovo mcp
0.00required app surface absent: rovo:mcp module
k llm model current
0.00vacuous — precondition unmet: >= 1 LLM call
k manifest semantics
0.00vacuous — precondition unmet: >= 1 function and >= 1 product call observed
k runtime risks
0.00vacuous — precondition unmet: >= 1 function and >= 1 product call observed
Issue updates flow trigger to queue to consumer: one ledger row per change under duplicate, reordered and dropped deliveries, multi-sprint changelog values, re-estimates followed, Retry-After honoured, no user context in background work.
t trigger handoff
0.00vacuous — precondition unmet: a trigger exists and >= 1 push observed
t event rows
0.00required app surface absent: trigger + consumer
t no double count
0.00vacuous — precondition unmet: >= 1 ledger row
t out of order
0.00permuted deliveries: 0/11 rows exact, 0/3 sprints with oracle numbers
t multi sprint parse
0.000/22 multi-id Sprint changes recorded exactly
t reestimate followed
0.000/2 re-estimated sprints show the oracle numbers
t retry after honoured
0.00vacuous — precondition unmet: the consumer-path fault fired
t no user in async
0.00vacuous — precondition unmet: >= 1 background Jira call
The first scheduled run backfills every change since each active sprint started, removals included; later runs heal what the event stream missed, page through results in each endpoint's own style, wait out rate limits and finish inside the module timeout.
r backfill complete
0.00required app surface absent: scheduledTrigger module
r removals found
0.000/7 removed-to-backlog changes found
r heal dropped
0.000/4 dropped changes healed exactly once; 0 other row(s) added
r pagination
0.00vacuous — precondition unmet: >= 1 read the site served in two or more pages
r rate limit
0.00vacuous — precondition unmet: the reconcile fault fired
r as app
0.00vacuous — precondition unmet: >= 1 scheduled-run Jira read
r completes in timeout
0.00vacuous — precondition unmet: a scheduled trigger made >= 1 Jira call
r idempotent rerun
0.00vacuous — precondition unmet: >= 1 ledger row
Forge KVS with the storage scope, ledger reads through the declared entity index in change-time order, no KVS limit errors.
s storage scope
0.00vacuous — precondition unmet: >= 1 KVS call
s entity index used
0.00vacuous — precondition unmet: >= 1 ledger read
s index order
0.00vacuous — precondition unmet: table rows rendered
s limits
0.00vacuous — precondition unmet: >= 1 KVS write
Every invoked resolver exists and answers structured errors, nobody sees changes to issues they cannot browse (not on screen, not in an LLM prompt, not in a Realtime payload), the hidden-change count is right, and each click posts exactly one ADF comment as the viewer.
b invoke contract
0.00vacuous — precondition unmet: >= 1 invoke observed
b no permission leak
0.00vacuous — precondition unmet: >= 1 person-facing change list returned or rendered
b hidden count
0.00required app surface absent: hidden-count / hiddenChanges
b comment adf as user
0.00vacuous — precondition unmet: >= 1 comment POST
b comment exactly once
0.00vacuous — precondition unmet: >= 1 comment POST
b realtime payload clean
0.00vacuous — precondition unmet: >= 1 realtime publish
The dashboard widget's numbers and chart, its board choice saved through the host, live updates through Forge Realtime, the sprint action's ledger table, sorting, issue links, comment flow, the Forge LLM explanation, close and not-started states.
u widget loadsdiagnostic, no weight
0.00required app surface absent: dashboards:widget view
u widget numbers
0.00required app surface absent: configured widget view
u widget chart
0.00required app surface absent: configured widget view
u widget edit config
0.00required app surface absent: dashboards:widget
u ledger table
0.00required app surface absent: jira:sprintAction
u ledger sort
0.00vacuous — precondition unmet: table rows rendered
u issue router
0.00vacuous — precondition unmet: table rows rendered
u comment flow
0.00required app surface absent: post-summary flow
u modal close
0.00vacuous — precondition unmet: modal rendered
u not started
0.00required app surface absent: jira:sprintAction
u widget live
0.00vacuous — precondition unmet: widget rendered sprints
u llm explain
0.00vacuous — precondition unmet: explain control rendered
Atlassian design tokens with 4.5:1 contrast in light and dark, a painted surface, no Content Security Policy violations or console errors, and the widget readable at 380 px wide.
v theme tokens
0.00vacuous — precondition unmet: a surface rendered app content
v dark mode
0.00vacuous — precondition unmet: a surface rendered app content
v csp clean
0.00vacuous — precondition unmet: a surface rendered app content
v console clean
0.00vacuous — precondition unmet: a surface rendered app content
v widget sizes
0.00vacuous — precondition unmet: widget rendered sprints
The get-sprint-scope action returns exact numbers and the visible changes for the invoking person, returns errors instead of throwing, and the skill's SKILL.md tells the agent how to use it.
a action result
0.00required app surface absent: action get-sprint-scope
a action errors
0.00vacuous — precondition unmet: the action exists
a action permissions
0.00required app surface absent: action get-sprint-scope
a skill instructions
0.00vacuous — precondition unmet: SKILL.md exists
Few Jira requests in the backfill and per event, few round trips before each surface paints, and a scheduled run with nothing new writes nothing — paid only in proportion to the excellence gate.
e reconcile economy
0.00mean over 3 scoring sites: backfill incomplete — economy is not credited on unfinished work
e event economy
0.00mean over 3 scoring sites: event rows inexact — economy is not credited on unfinished work
e ui round trips
0.00mean over 3 scoring sites: no surface rendered app content
Graded browser recording
No recording: no app surface ever rendered.
v_theme_tokens, v_dark_mode, v_csp_clean, v_console_clean: vacuous — precondition unmet: a surface rendered app content
Run details
- Model
- nvidia/nemotron-3.5-lightning
- Engine events
- 0
- Repair rounds
- 0
- Started
- Oct 6, 2026, 03:07 AM
- Finished
- Oct 6, 2026, 03:14 AM