▣
Forecast Commit Classifier
SalesPaste a list of open deals with stage, amount, close date and notes; get each bucketed into Commit / Best-Case / Pipeline / Omit with a one-line rationale, plus a weighted roll-up, a single recommended forecast number, and the top risks that could break it.
The routing recipe
Job typeComplex reasoningPolicyBest qualityConstraintsctx ≥ 128k
routes to
NVIDIA Nemotron 3 Ultra 550B-A55B nvidia
nvidia:nvidia/nemotron-3-ultra-550b-a55b
93.7
match score
96.5
task fit
100
cost
59.5
speed
Why this model from benchmarks & capabilities
- "Complex reasoning" leans hardest on reasoning; NVIDIA Nemotron 3 Ultra 550B-A55B scores 100/100 there — top-1 of 54 chat models.
- Strength on this class of work shows up on GPQA Diamond, MMLU-Pro, BIG-Bench Hard, where NVIDIA models are competitive.
- Cost: runs on NVIDIA free credits / self-hostable NIM — effectively $0 marginal. Speed: ~70 tok/s (estimated).
- Chosen over DeepSeek V4 Pro for cheaper & faster.
| Capability the job needs | Weight | NVIDIA Nemotron 3 Ultra 550B-A55B | Catalog rank |
|---|---|---|---|
| reasoning | 1.00 | #1 of 54 | |
| math | 0.60 | #2 of 54 | |
| knowledge | 0.50 | #3 of 54 |
Relevant benchmarksGPQA Diamond MMLU-Pro BIG-Bench Hard LiveBench (reasoning)
EconomicsCost: runs on NVIDIA free credits / self-hostable NIM — effectively $0 marginal.
Speed: ~70 tok/s (estimated). vs runner-upChosen over DeepSeek V4 Pro for cheaper & faster. (DeepSeek V4 Pro, score 91.8)
Speed: ~70 tok/s (estimated). vs runner-upChosen over DeepSeek V4 Pro for cheaper & faster. (DeepSeek V4 Pro, score 91.8)
Example result
Sample input ▾
Forecast list — Q3 close, AE: Marcus Lee (iCompaas). Today 2026-06-29.
1. Acme Fintech — Stage: Contracting — $72k — close Jul 8 — CISO + procurement engaged, redlines with legal, signature target Jul 8, PCI audit deadline Aug.
2. Northwind Health — Stage: Negotiation — $48k — close Jul 31 — single contact (analyst), no buyer, ghosted 11 days, close already slipped once.
3. Bluepeak SaaS — Stage: Proposal — $36k — close Jul 22 — VP Eng champion, verbal yes, no procurement yet, displacing Vanta, demo of evidence automation landed.
4. Cobalt Logistics — Stage: Demo — $60k — close Jul 30 — strong discovery, ISO 27001 mandate from new EU customer, but only met IT manager, no budget convo.
5. Drake Robotics — Stage: Discovery — $25k — close Jul 15 — one call 5 weeks ago, no reply since, no next step.
PER-DEAL TABLE:
| Deal | Amount | Stage | Category | Rationale |
|---|---|---|---|---|
| Acme Fintech | $72k | Contracting | COMMIT | Economic buyer (CISO) + procurement engaged, in redlines, dated signature Jul 8, Aug PCI audit = hard compelling event. All commit ingredients present. |
| Bluepeak SaaS | $36k | Proposal | BEST-CASE | Champion verbal yes and strong product fit displacing Vanta, but missing ingredient: procurement/paper not started and no economic buyer met. |
| Cobalt Logistics | $60k | Demo | PIPELINE | Real ISO 27001 mandate with a dated EU-customer driver, but single-threaded to an IT manager and no budget conversation — upside, not this period. |
| Northwind Health | $48k | Negotiation | OMIT | Single-threaded to a junior analyst, no economic buyer, ghosted 11 days, close already slipped once — exclude despite 'Negotiation' stage. |
| Drake Robotics | $25k | Discovery | OMIT | One call 5 weeks ago, no reply, no next step — dead until re-engaged. |
ROLL-UP: gross pipeline $241k | weighted forecast $108.0k (Acme 0.9×72=64.8 + Bluepeak 0.6×36=21.6 + Cobalt 0.25×60=15.0 + 0) | Commit total $72k | Best-Case total $36k
THE CALL: Forecast $72k Commit with a realistic stretch to $108k. Confidence: High on the $72k (one clean, buyer-backed deal with a dated event), Medium on the stretch — it hangs almost entirely on Bluepeak converting verbal to paper.
TOP 3 RISKS TO THE NUMBER:
1. Concentration risk — 60% of the Commit number is a single deal (Acme). De-risk: confirm Jul 8 signature in writing this week and watch for legal-redline drift on the DPA.
2. Bluepeak is a verbal, not a commit — no buyer, no paper. De-risk: get the VP Eng to introduce procurement and start the order form now; until paper moves it cannot be upgraded.
3. Cobalt is the most likely to be over-forecast — a genuine ISO mandate makes it tempting to pull forward. De-risk: hold at Pipeline until the budget owner is met; a dated mandate with no budget contact is not a Q3 deal yet.
Test it on your own data
\ntop 3 risks to the number: ","task":"reasoning","optimize":"best_quality","mincontext":128000}'>