Head to head / curated pair / cycle 04
Convex Chef vs Totalum
Both products were given the identical 9 prompt vibeOps specification, 5 times each. Nothing on this page is an impression, all of it is a measurement from the same harness.
The reactive backend underneath it does most of the work the operational prompts ask for, so scheduled functions, durable actions and consistent reads arrived without prompting. That is why it posts one of the best API subscores in the cohort despite a middling agent score. Where it struggles is anything conventionally relational: prompt 2 needed rework in three of five runs to express the tenancy constraint the way the platform prefers.
The most complete machine facing surface we measured: a real REST layer, signed webhooks with a retry ladder and an MCP server that a client could enumerate without extra work. Public pages came back server rendered, which is why it posts the largest no JavaScript payload in the cohort. Design output is functional rather than expressive and the agent needed more explicit prompting than the top scorer on axis one.
Index gap
17.46
Totalum leads
Axes won
1 / 9
Convex Chef versus Totalum, of ten axes
Faster median
Convex Chef
88.7 s per prompt
Higher pass rate
Totalum
93.3%
Every measurement, side by side
Cyan marks the better figure on each row. The delta column is the first product minus the second.
| Measurement | Convex Chef | Totalum | Delta |
|---|---|---|---|
| Composite index | 70.50 | 87.96 | -17.46 |
| Agent performance (weight 18) | 75.0 | 94.0 | -19.00 |
| Reliability (weight 18) | 60.0 | 93.0 | -33.00 |
| Scalability (weight 13) | 84.0 | 92.0 | -8.00 |
| SEO and GEO (weight 12) | 65.0 | 95.0 | -30.00 |
| API and MCP (weight 9) | 82.0 | 97.0 | -15.00 |
| Integrations (weight 8) | 72.0 | 88.0 | -16.00 |
| Design output (weight 7) | 76.0 | 86.0 | -10.00 |
| Speed (weight 7) | 35.4 | 29.5 | +5.90 |
| Value (weight 5) | 78.0 | 84.0 | -6.00 |
| Code ownership (weight 3) | 88.0 | 96.0 | -8.00 |
| Median seconds per prompt | 88.7 | 106.3 | -17.60 |
| p10 seconds | 51.5 | 59.2 | -7.70 |
| p90 seconds | 142.4 | 167.6 | -25.20 |
| Executions failed | 10 | 1 | +9.00 |
| HTML bytes, no JS | 12.0 KB | 41.0 KB | -29696.00 |
| LCP milliseconds | 1,840 | 1,480 | +360.00 |
| CLS | 0.040 | 0.020 | +0.02 |
| Entry price EUR | 25.00 | 29.00 | -4.00 |
Screenshot diff
Both landing pages, captured by the lab on 14 Aug 2026 at 1440 by 900. Stored locally, never hotlinked.
Convex Chef, captured 14 Aug 2026.
Totalum, captured 14 Aug 2026.
Speed spread on a shared scale
Convex Chef
51.5 / 88.7 / 142.4 seconds
Totalum
59.2 / 106.3 / 167.6 seconds
Axis by axis, who wins
Convex Chef wins 1 of 10
- Speed35.4
Totalum wins 9 of 10
- Agent performance94.0
- Reliability93.0
- Scalability92.0
- SEO and GEO95.0
- API and MCP97.0
- Integrations88.0
- Design output86.0
- Value84.0
- Code ownership96.0
Verdict
Totalum takes the composite by 17.46 points. It wins on the weighted total, but the axis table is where the real decision sits: Convex Chef takes 1 axes and Totalum takes 9.
If the work in front of you is mostly interface, weight the design and speed rows. If it is a product with an API, webhooks and background jobs, weight agent performance, reliability and the API and MCP row, which together carry 45 of the 100 index points. If the project has to outlive its vendor, read the code ownership row first and treat everything else as secondary.
Both products in this comparison were measured in the same cycle, on the same specification, by the same harness. The per run data for each is on its product page and in /api/runs.json.