Update README.md
Browse files
README.md
CHANGED
|
@@ -82,23 +82,23 @@ Welcome to try out our online demo [here](https://dr.miromind.ai/). In this demo
|
|
| 82 |
|
| 83 |
### More Benchmarks
|
| 84 |
|
| 85 |
-
| Method | HLE<br>Pass@1 | Frames<br>Pass@1 | BrowseComp<br>Pass@1 | BrowseComp-ZH
|
| 86 |
-
|
| 87 |
-
| OpenAI Deep Research | 26.6 | - | 51.5 |
|
| 88 |
-
| Gemini Deep Research | 26.9 | - | - |
|
| 89 |
-
| Kimi-Researcher | 26.9 | 78.8 | - |
|
| 90 |
-
| | | | |
|
| 91 |
-
| WebDancer-7B | - | - | - |
|
| 92 |
-
| WebSailor-7B | - | - | 6.7 |
|
| 93 |
-
| MiroThinker-8B-SFT-v0.1 | - | 58.0 | 5.5 |
|
| 94 |
-
| MiroThinker-8B-DPO-v0.1 | - | 64.4 | 8.7 |
|
| 95 |
-
| | | | |
|
| 96 |
-
| WebThinker-32B-RL | - | - | - |
|
| 97 |
-
| WebDancer-QwQ-32B | - | - | 3.8 |
|
| 98 |
-
| WebSailor-32B | - | - | 10.5 |
|
| 99 |
-
| WebShaper-32B | - | - | - |
|
| 100 |
-
| MiroThinker-32B-SFT-v0.1 | 10.2 | 70.4 | 10.6 |
|
| 101 |
-
| <span style="white-space:nowrap;">MiroThinker-32B-DPO-v0.1</span> | 11.8 | 71.7 | 13.0 |
|
| 102 |
|
| 103 |
1. MiroThinker’s performance was tested with [this repository](https://github.com/MiroMindAI/MiroThinker) and open-source tools; other models’ results are from their papers and official sites.
|
| 104 |
|
|
|
|
| 82 |
|
| 83 |
### More Benchmarks
|
| 84 |
|
| 85 |
+
| Method | HLE<br>Pass@1 | Frames<br>Pass@1 | BrowseComp<br>Pass@1 | <span style="white-space:nowrap;">BrowseComp-ZH</span><br>Pass@1 | WebWalkerQA<br>Pass@1 |
|
| 86 |
+
|-------------------------------------------------------------------|:-------------:|:----------------:|:--------------------:|:----------------------------------------------------------------:|:---------------------:|
|
| 87 |
+
| OpenAI Deep Research | 26.6 | - | 51.5 | 42.9 | - |
|
| 88 |
+
| Gemini Deep Research | 26.9 | - | - | - | - |
|
| 89 |
+
| Kimi-Researcher | 26.9 | 78.8 | - | - | - |
|
| 90 |
+
| | | | | | |
|
| 91 |
+
| WebDancer-7B | - | - | - | - | 36.0 |
|
| 92 |
+
| WebSailor-7B | - | - | 6.7 | 14.2 | - |
|
| 93 |
+
| MiroThinker-8B-SFT-v0.1 | - | 58.0 | 5.5 | 9.3 | 41.3 |
|
| 94 |
+
| MiroThinker-8B-DPO-v0.1 | - | 64.4 | 8.7 | 13.6 | 45.7 |
|
| 95 |
+
| | | | | | |
|
| 96 |
+
| WebThinker-32B-RL | - | - | - | - | 46.5 |
|
| 97 |
+
| WebDancer-QwQ-32B | - | - | 3.8 | 18.0 | 47.9 |
|
| 98 |
+
| WebSailor-32B | - | - | 10.5 | 25.5 | - |
|
| 99 |
+
| WebShaper-32B | - | - | - | - | 51.4 |
|
| 100 |
+
| MiroThinker-32B-SFT-v0.1 | 10.2 | 70.4 | 10.6 | 13.8 | 45.7 |
|
| 101 |
+
| <span style="white-space:nowrap;">MiroThinker-32B-DPO-v0.1</span> | 11.8 | 71.7 | 13.0 | 17.0 | 49.3 |
|
| 102 |
|
| 103 |
1. MiroThinker’s performance was tested with [this repository](https://github.com/MiroMindAI/MiroThinker) and open-source tools; other models’ results are from their papers and official sites.
|
| 104 |
|