Very long inputs
Choose Mistral Small 4
Mistral Small 4 reads up to 256k tokens per request, against 8k for Tiny Aya Global.
Mistral Small 4 vs Tiny Aya Global: price, context and capabilities
Choose Mistral Small 4 for very long inputs, images as input and the latest release; on the other points the two are level or the data is missing.
In short
- A price comparison is not possible: one of them has no comparable published price.
- Mistral Small 4 reads more per request: 256k tokens against 8k.
- Mistral Small 4 is the newer model (March 2026, against Feb 17, 2026).
- Only Mistral Small 4 offers vision, reasoning, tool calling, structured output.
Which should you choose?
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Images as input
Choose Mistral Small 4
Mistral Small 4 accepts images as input; Tiny Aya Global does not.
Latest release
Choose Mistral Small 4
Mistral Small 4 is the newer release (March 2026, against Feb 17, 2026 for Tiny Aya Global).
Specifications side by side
| Mistral Small 4 | Tiny Aya Global | |
|---|---|---|
| Provider | Mistral AI | Cohere |
| Category | Language models | Language models |
| Status | Generally available | Generally available |
| Headline price | No data | No data |
| Input | No data | No data |
| Output | No data | No data |
| Cached input | No data | No data |
| Context window | 256k | 8k |
| Max output | No data | No data |
| Released | March 2026 | Feb 17, 2026 |
| Knowledge cutoff | No data | No data |
| Licence | Open weights | Open weights |
| Capabilities | Vision, Reasoning, Tool calling, Structured output, Self-hostable | Self-hostable |
| API model id | mistral-small-2603 | tiny-aya-global |
Benchmarks side by side
Leader in bold.
| Benchmark | Mistral Small 4 | Tiny Aya Global |
|---|---|---|
| GPQA Diamond | 76.9% | 30.5% |
| Humanity's Last Exam | 9.9% | 5.2% |
| IFBench (instruction following) | 48.2% | 20.1% |
| Intelligence Index | 11.3 | No data |
| Terminal-Bench Hard | 17.4% | No data |
| SciCode | 38.8% | No data |
| τ²-Bench (agentic tool use) | 41.2% | No data |
| MMMU-Pro (visual reasoning) | 56.8% | No data |
| Long-context reasoning (AA-LCR) | 49.7% | No data |
Independent evaluations by Artificial Analysis, run the same way for every model, at the reasoning setting shown. Percentages unless marked as an index.
Frequently asked questions
Should I choose Mistral Small 4 or Tiny Aya Global?
Choose Mistral Small 4 for very long inputs, images as input and the latest release; on the other points the two are level or the data is missing.
Which is cheaper, Mistral Small 4 or Tiny Aya Global?
A price comparison is not possible: one of them has no comparable published price.
Which has the larger context window, Mistral Small 4 or Tiny Aya Global?
Mistral Small 4 reads more per request: 256k tokens against 8k.
What is the difference between Mistral Small 4 and Tiny Aya Global?
Mistral Small 4 is made by Mistral AI and Tiny Aya Global by Cohere. A price comparison is not possible: one of them has no comparable published price. Mistral Small 4 reads more per request: 256k tokens against 8k. Mistral Small 4 is the newer model (March 2026, against Feb 17, 2026). Only Mistral Small 4 offers vision, reasoning, tool calling, structured output.
Which is newer, Mistral Small 4 or Tiny Aya Global?
Mistral Small 4 was released on March 2026 and Tiny Aya Global on Feb 17, 2026.