# Is Mage-VL-Cowork SOTA?

Pending benchmarks. Announced Oct 30, 2026; no independent results yet.

As of Nov 3, 2026, Mage-VL-Cowork is pending benchmarks on isitsota: announced Oct 30, 2026, with no independent results on the boards we track yet.

Microsoft's 4B open-weight computer-use agent. One vision-language policy works a desktop either through screenshot-grounded GUI actions or through a persistent bash and Python shell, choosing the channel turn by turn. Fine-tuned from an intermediate Mage-VL checkpoint; Apache-2.0.

Canonical: https://isitsota.com/m/mage-vl-cowork
Last updated: 2026-10-30
Tracking since: 2026-10-30

## Announced facts

| Fact | Value |
| --- | --- |
| Lab | Microsoft |
| Announced | Oct 30, 2026 |
| API | Not yet |
| Open weights | Published |
| Parameters | 4.74B stored (BF16 safetensors); Microsoft lists it as 4B |
| Type | Hybrid computer-use agent: image + text in, text out (think rationale, action summary and a tool call) |
| Base model | microsoft/Mage-VL (intermediate 4B checkpoint) |
| Backbone | Mage-ViT vision encoder + Qwen3-4B-Instruct-2507 (llava_onevision2 architecture) |
| Action channels | computer_use (11 GUI primitives on normalized [0, 999] coordinates) and bash (persistent shell and Python environment) |
| Context | 32,768 tokens used in training and inference; architecture supports up to 262,144 |
| Training | Five stages: GUI midtraining, GUI agent SFT, coding and terminal SFT, hybrid SFT, then online GRPO on 1,988 verifiable tasks; 32 NVIDIA B200 GPUs |
| License | Apache-2.0 |
| Intended use | Released for research purposes, not for product or service deployment (Mage repo notice) |
| Release | Weights on Hugging Face Oct 30, 2026; project page and README were committed Oct 27–28 without weights |

Tags marked unsure are not stated on the lab's own pages.

## What would let us call it

Verdicts use only independent boards we already track. Two consecutive snapshots are required before any crown.

- [x] Publicly announced with a source
- [x] Usable by the public (API or weights)
- [ ] Listed on an independent board we track
- [ ] Enough results for a rank in a category
- [ ] Holds across two snapshots (crowns only)

## What the lab claims

Self-reported. Shown for context; not used in any verdict.

- Lab claim: 66.0% on OSWorld-Verified (two-run average over 361 tasks excluding Google Drive, 50-step budget, Microsoft's MAGEHybridAgent harness), with 87.5% on OS tasks and 62.7% on multi-application workflows.
- Lab claim: online RL adds 10.16 points over the hybrid SFT stage while cutting model calls per task by 13.0% and output tokens by 25.4%.

## Sources

- REPO: [microsoft/Mage-VL-Cowork model card (Hugging Face)](https://huggingface.co/microsoft/Mage-VL-Cowork) (2026-10-30)
- LAB: [Mage-VL-Cowork project page](https://microsoft.github.io/Mage/cowork/)
- LAB: [Mage-VL-Cowork technical report (PDF)](https://github.com/microsoft/Mage/blob/main/assets/mage_vl_cowork_tech_report.pdf) (2026-10-30)
- REPO: [Mage-VL-Cowork code (GitHub)](https://github.com/microsoft/Mage/tree/main/mage_vl_cowork)

Tracking continues until an independent board lists it.

## Related

- Lab: [Microsoft](https://isitsota.com/labs/microsoft.md)
- All pending releases: https://isitsota.com/releases.md
- JSON: https://api.isitsota.com/v1/models/mage-vl-cowork