Cohere vs Thinking Machines
Verdict
- Thinking Machines's flagship leads: Inkling Small at 55.0 vs Command A+ at 54.0.
- Thinking Machines has the better model in 3 of 4 price bands.
- Cohere's fastest ranked model (Command A+, 236 tok/s) beats Thinking Machines's (Inkling Small, 133 tok/s).
- Cohere publishes more open-weights models (13 vs 1).
Best model per price band
Each provider’s highest-index model below the price ceiling. Purple marks the winner of the band.
| Price band | Cohere | Thinking Machines |
|---|---|---|
| Under $0.50/M | – | – |
| Under $2/M | – | Inkling Small 55.0$0.638/M |
| Under $10/M | Command A 41.7$4.38/M | Inkling Small 55.0$0.638/M |
| Any price | Command A 41.7$4.38/M | Inkling Small 55.0$0.638/M |
Line-up
| Metric | Cohere | Thinking Machines |
|---|---|---|
| Flagship index | 54.0 | 55.0 |
| Ranked models | 6 | 1 |
| Open-weights models | 13 | 1 |
| Cheapest ranked, $/M | $4.38 | $0.638 |
| Fastest ranked, tok/s | 236 | 133 |
| Release cadence, days | 21 | – |
| Last release | 9 Jun 2026 | – |
| Next expected | 21 Jun 2026 – 8 Jul 2026 | – |
Data as of 10 Sept 2026. See Cohere and Thinking Machines in full.