Models22in catalog
Largest context131KDolphin 2 9 2 Qwen2 72b
Avg context
75K
Cheapest input
$0.070/M
Phi 4
Largest output
131K
Dolphin 2 9 2 Qwen2 72b
Speed tiers
deep 2fast 8balanced 12
Capabilities
What Microsoft models support across the catalog.
Vision
2 / 22
9% of models
Tool use
5 / 22
23% of models
Function calling
5 / 22
23% of models
Extended thinking
1 / 22
5% of models
Streaming
22 / 22
100% of models
Prompt caching
1 / 22
5% of models
All Microsoft models
Sorted by context window, largest first. Tap a row for full specs.
| Model | Context | Max output | Input $/M | Speed |
|---|---|---|---|---|
| Dolphin 2 9 2 Qwen2 72bQwen | 131K | 131K | $0.900/M | deep |
| Phi 4 Mini | 131K | 4K | $0.075/M | fast |
| Phi 4 Mini Reasoning | 131K | 4K | $0.080/M | fast |
| Phi 4 Multimodal | 131K | 4K | $0.080/M | balanced |
| Phi 3 5 Mini | 128K | 4K | $0.130/M | fast |
| Phi 3 5 Moe | 128K | 4K | $0.160/M | balanced |
| Phi 3 5 Vision | 128K | 4K | $0.130/M | balanced |
| Phi 3 Medium 128k | 128K | 4K | $0.170/M | balanced |
| Phi 3 Mini 128k | 128K | 4K | $0.100/M | fast |
| Phi 3 Small 128k | 128K | 4K | $0.150/M | balanced |
| Phi 4 Mini Instruct | 128K | 128K | — | fast |
| WizardLM-2 8x22B | 66K | 66K | $0.480/M | balanced |
| Openorca 7b | 33K | 33K | $0.200/M | fast |
| Phi 4 Reasoning | 33K | 4K | $0.125/M | deep |
| Phi 3 Vision 128k | 32K | 32K | $0.200/M | balanced |
| Chatdolphin | 16K | 16K | $0.500/M | balanced |
| Dolphin | 16K | 16K | $0.500/M | balanced |
| Phi 4 | 16K | 16K | $0.070/M | balanced |
| Phi 3 Small 8k | 8K | 4K | $0.150/M | balanced |
| Phi 3 Medium 4k | 4K | 4K | $0.170/M | balanced |
| Phi 3 Mini 4k | 4K | 4K | $0.130/M | fast |
| Phi 2 3b | 2K | 2K | $0.100/M | fast |
See rankings
How Microsoft models rank against the full catalog.