GPT-6.1 Sol vs Mistral Small 3.2 24B Instruct (Venice)
On the 26 use cases both models ran, GPT-6.1 Sol won 23 to 3 (better outcome first, then lower cost).
Pass rate: GPT-6.1 Sol 100% (26 runs) vs Mistral Small 3.2 24B Instruct (Venice) 12% (26 runs). Average cost per run: $0.05 vs $0.0043.
| Metric | GPT-6.1 Sol | Mistral Small 3.2 24B Instruct (Venice) |
|---|---|---|
| Runs on shared use cases | 26 | 26 |
| Pass rate | 100% (26/26) | 12% (3/26) |
| Avg cost per run | $0.05 | $0.0043 |
| Avg wall time | 1m | 2m |
| Step-ins per run | 0.00 | 0.00 |
| Use cases won | 23 | 3 |
Use case by use case
- AIS collision risk: closest point of approach and alert levels: GPT-6.1 Sol better (pass 100% vs 0%)
- Anchor drag alarm from a GPS track: GPT-6.1 Sol better (pass 100% vs 0%)
- Backtest a moving-average crossover on a bundled synthetic price file: GPT-6.1 Sol better (pass 100% vs 0%)
- Best fishing windows from tide, solunar and weather data: GPT-6.1 Sol better (pass 100% vs 0%)
- Best sailing angle and leg times from a polar table: Mistral Small 3.2 24B Instruct (Venice) better (pass 100% vs 100%)
- Boat fuel range with a reserve over a planned trip: GPT-6.1 Sol better (pass 100% vs 0%)
- Build a TV home screen that shows a smart home's device status instead of ads: GPT-6.1 Sol better (pass 100% vs 0%)
- Catch log checked against a made-up fishing regulations table: GPT-6.1 Sol better (pass 100% vs 0%)
- Debt avalanche vs snowball schedule to the cent: GPT-6.1 Sol better (pass 100% vs 0%)
- Discounted cash flow valuation with mid-year option and a sensitivity grid: Mistral Small 3.2 24B Instruct (Venice) better (pass 100% vs 100%)
- Find your own listings on people-search pages and draft opt-out requests for approval: GPT-6.1 Sol better (pass 100% vs 0%)
- Go/no-go departure time from a gridded wind file: GPT-6.1 Sol better (pass 100% vs 0%)
- Model a game-ready low-poly 3D prop from a one-line brief, within a tight triangle budget: GPT-6.1 Sol better (pass 100% vs 0%)
- Mortgage amortization with extra monthly payments and lump sums: Mistral Small 3.2 24B Instruct (Venice) better (pass 100% vs 100%)
- Pension claiming-age comparison: lifetime totals, present values and break-even ages: GPT-6.1 Sol better (pass 100% vs 0%)
- Progressive tax on fictional brackets, and pre-tax vs after-tax savings comparison: GPT-6.1 Sol better (pass 100% vs 0%)
- Read a synthetic annual-report excerpt and compute financial ratios: GPT-6.1 Sol better (pass 100% vs 0%)
- Rebalance to target weights with whole shares and highest-cost-first tax lots: GPT-6.1 Sol better (pass 100% vs 0%)
- Required minimum distributions from a supplied divisor table, with and without delaying the first one: GPT-6.1 Sol better (pass 100% vs 0%)
- Retirement drawdown with a 4% rule, inflation raises and a skip-raise-after-loss guardrail: GPT-6.1 Sol better (pass 100% vs 0%)
- Sailboat instrument dashboard from an NMEA 0183 log: GPT-6.1 Sol better (pass 100% vs 0%)
- Seeded Monte Carlo retirement odds against a stated success threshold: GPT-6.1 Sol better (pass 100% vs 0%)
- Tide height by rule of twelfths and cosine, with under-keel clearance windows: GPT-6.1 Sol better (pass 100% vs 0%)
- Triage a work inbox into read-only summaries and reply drafts, sending nothing: GPT-6.1 Sol better (pass 100% vs 0%)
- Water-quality sensor anomaly flags and daily means: GPT-6.1 Sol better (pass 100% vs 0%)
- Work backward from a retirement budget to a savings target, contributions and earliest retirement age: GPT-6.1 Sol better (pass 100% vs 0%)
Only use cases both models ran are compared; all settings of each model are grouped. How we score.