ChatGPT
- Recorded model statements
- 2 models evidenced as usable1 rolling out
- Frontier models
- evidenced: GPT-6 Astra (limited availability)
- Adds
- no statement recorded
- Connects to
- no statement recorded
- Last change
- —
- Sources
Compare
ChatGPT vs Microsoft 365 Copilot
Two to four products side by side: what is recorded for each today, and where their statements differ. Where a product has no statement, that is said — it is not a difference. There is no winner here.
Microsoft 365 Copilotmodels verified for the Cowork area only
- Recorded model statements
- 7 models evidenced as usable3 statements not verified1 × Model retired
- Frontier models
- evidenced: GPT-6 Astra (limited availability) · Claude Fable 5.1 (limited availability)
- Adds
- Agent mode (limited availability) · Deep research
- Connects to
- no statement recorded
- Last change
- 5 months ago
- Sources
7learn.microsoft.com+2
- learn.microsoft.com/en-us/microsoft-365/copilot/cowork/cowork-models
- www.microsoft.com/en-us/copilot/blog/2025/09/24/expanding-model-choice-in-microsoft-365-copilot/
- www.microsoft.com/en-us/copilot/blog/2025/08/07/available-today-gpt-5-in-microsoft-365-copilot/
- www.microsoft.com/en-us/copilot/blog/2025/12/11/available-today-gpt-5-2-in-microsoft-365-copilot/
- techcommunity.microsoft.com/blog/microsoft-copilot-blog/available-today-gpt-5-4-thinking-in-microsoft-365-copilot/4499746
- www.microsoft.com/en-us/copilot/blog/2025/09/29/vibe-working-introducing-agent-mode-and-office-agent-in-microsoft-365-copilot/
- www.microsoft.com/en-us/copilot/blog/2025/06/02/researcher-and-analyst-are-now-generally-available-in-microsoft-365-copilot/
What can be compared
Only what at least two of the products have a verified statement for is compared here. Solid: available. Light: limited. Ring: rolling out. Where a product has no statement, the cell says “no statement” — which is not “not available”. Below each group: the statements recorded per product, as plain counts. Model availability is checked against the provider’s pages, and a statement without a check reads “not verified”; capabilities and integrations are recorded from sourced updates.
| Compare | ChatGPT | Microsoft 365 Copilotmodels verified for the Cowork area only |
|---|---|---|
| Models (2 comparable) | ||
| GPT-5.6 Sol | Limited availability | Limited availability |
| GPT-6 Astra | Limited availability | Limited availability |
| recorded statements | 2 models evidenced as usable1 rolling out | 7 models evidenced as usable3 statements not verified1 × Model retired |
| Capabilities (0 comparable) | ||
| No entry has a verified statement for two of these products. | ||
| recorded statements | no statement recorded | 2 capabilities evidenced as usable |
Where they differ
Where at least two of the products have a verified statement, nothing differs.
What they share
2 entries have the same status in every product that has a verified statement for them.
Show dates and sources — What they share
| Compare | ChatGPT | Microsoft 365 Copilotmodels verified for the Cowork area only |
|---|---|---|
| Model · Models (2) | ||
| GPT-5.6 Sol | Limited availabilitystart not documented | Limited availabilitystart not documented |
| GPT-6 Astra | Limited availabilitystart not documented | Limited availabilitystart not documented |
What cannot be compared
For 12 entries at most one of the products has a verified statement. That is not a difference: for the others there is no statement, or one that was not checked.
Show dates and sources — What cannot be compared
ChatGPT: 1
Models GPT-5.6 Luna (rolling out)
Microsoft 365 Copilot: 11
Models Claude Fable 5.1 (limited availability) · Claude Opus 4.1 (model retired) · Claude Opus 5 (limited availability) · Claude Sonnet 5 (limited availability) · GPT-5 (not verified) · GPT-5.2 (not verified) · GPT-5.4 (not verified) · GPT-5.5 (limited availability) · GPT-5.6 Terra (limited availability)
Capabilities Agent mode (limited availability) · Deep research
No winneragi.jetzt compares facts. There is no overall score, no “best”, no recommendation. agi.jetzt does not test Work AI products. Nothing here is a performance result — only documented availability and features.