Lab vs Lab
Cohere vs OpenAI.
The club-level face-off: Club Scores, best models, squad depth, division titles and the price floor — every number computed from the model database.
| Cohere | OpenAI | Edge | |
|---|---|---|---|
| Club Score | 0.0 | 94.4 | OpenAI |
| Best model | Command A+ | GPT-6 Astra | OpenAI |
| Top-3 average | 0.0 | 90.9 | OpenAI |
| Squad size | 2 | 6 | OpenAI |
| Cheapest capable ($/1M blended) | 4.4 | 0.5 | OpenAI |
| Open-weight share % | 100 | 0 | Cohere |
| Division titles | 0 | 1 | OpenAI |
OpenAI holds the #1 model in: Reasoning Division.
Price floor — cheapest capable model on three workloads
| Workload | Cohere | OpenAI |
|---|---|---|
| Chat assistant (1M in / 250K out) | $5.00 | $0.50 |
| Document analysis (5M in / 500K out) | $17.50 | $1.60 |
| Code generation (2M in / 1000K out) | $15.00 | $1.60 |
Best vs best
League fixture: The two clubs' #1 models, head-to-head. In Season Zero, fixture results are decided by the data — the category winners above. When voted fixtures launch, results blend 60% benchmark / 40% crowd per the published rulebook.
Verdict by category
| Category | Command A+ | GPT-6 Astra | Winner |
|---|---|---|---|
| League Score | 0.0 | 96.7 | GPT-6 Astra |
| Coding (SWE-bench %) | 51.0 | — | — |
| Reasoning (GPQA %) | — | 96.0 | — |
| Context (tokens) | 128000 | 1050000 | GPT-6 Astra |
| Speed (tok/s) | — | — | — |
Tale of the tape
| Spec | Command A+ | GPT-6 Astra |
|---|---|---|
| Maker | Cohere | OpenAI |
| Released | May 20, 2026 | Sep 3, 2026 |
| League Score | 0 | 97 |
| Context | 128,000 tokens | 1,050,000 tokens |
| Max output | 64,000 | 128,000 |
| $/1M input | — | $10.00 |
| $/1M output | — | $50.00 |
| Weights | Open | Closed |
| Speed | — | — |
Head-to-head benchmarks (shared only)
| Benchmark | Command A+ | GPT-6 Astra |
|---|---|---|
| No shared benchmark measurements yet — shown as not measured, never interpolated. | ||
Cost at three workloads
| Workload | Command A+ | GPT-6 Astra |
|---|---|---|
| Light — 1M in / 250K out | — | $22.50 |
| Standard — 10M in / 2.5M out | — | $225.00 |
| Long-context — 50M in / 5M out | — | $750.00 |
The pick
Pick Command A+ if you want downloadable weights / self-hosting; coding is the priority.
Pick GPT-6 Astra if you want the higher overall League Score; you need the bigger context window.
All figures from the model database — sources and dates on each model page. Cost your own workload or compare a different pair.