OpenCSG Agentic-30B-A3B
ไธบ Skillsใๅทฅๅ
ท่ฐ็จๅ็งๆๅๆง่ก่่ฎญ็ป็ๆบ่ฝไฝๆจกๅ
An agent model trained for Skills, tool use, and private execution
OpenCSG ๆจกๅไปๅบ ยท ไฝฟ็จๅ่ฎฎ / License Agreement
ๆจกๅ็ฎไป
OpenCSG Agentic-30B-A3B ๆฏ OpenCSG Agentic ็ณปๅ็้ฆไธช็ๆฌใๆไปฌไปไฝๆฟๆดปๅๆฐ MoE ๆจกๅ่ตทๆญฅ๏ผๅด็ปๆฌๅฐๆบ่ฝไฝใไผไธ็งๆๅ้จ็ฝฒๅไธชไบบ็ป็ซฏไธ็ Skills ๆง่กๅๅ่ฎญ็ปใๅฎๆฒกๆ่ตฐโๅ ่ฎญ็ป่ถ ๅคง้็จๆจกๅ๏ผๅๅ็ผฉๆ่พๅฐๅฏๆฌโ็่ทฏ็บฟใ
่ฟไธช็ณปๅ่งฃๅณ็ๆฏไธไธชๅพๅ ทไฝ็้ฎ้ข๏ผๅฝๆบ่ฝไฝ็ๆญฃ่ฟๅ ฅ็จๆท็ฏๅข๏ผๅฎ่ฝๅฆ็่งฃๆไปคใ้ๆฉๆญฃ็กฎ็ Skillใ็ๆๅๆณๅๆฐใ่ฏปๅๅทฅๅ ท็ปๆ๏ผๅนถๆ่ทจๅคไธช็ณป็ป็ไปปๅกๅๅฎ๏ผๆจกๅๅจๅฏน่ฏไธญโ็ฅ้ๅพๅคโๅนถไธ็ญไบๅฎ่ฝ็จณๅฎๆไฝๆฅๅใ้ฎไปถใๅทฅๅใ็ฅ่ฏๅบใCRMใๅบๅญๆ่ชๅจๅไปปๅกใAgentic ็ณปๅๆๅ่ ไฝไธบไธป่ฆไผๅ็ฎๆ ใ
ๅฝๅ็ๆฌๅ ฑๆ 30.53B ๅๆฐ๏ผๆฏไธช token ็บฆๆฟๆดป 3.3B ๅๆฐใๅฎ้็จ 128 ไธชไธๅฎถใๆฏไธช token ๆฟๆดป 8 ไธชไธๅฎถ็ MoE ๆถๆใๆฌๆจกๅไปฅ BF16 ็ฒพๅบฆๅๅธ๏ผๅทฒๅจๅๅผ NVIDIA A800 80GB ไธๅฎๆ็ฌ็ซ้จ็ฝฒ๏ผvLLM ๅ ่ฝฝๆ้ๅ ็จ 56.88 GiB ๆพๅญ๏ผๅบๅฎ็ๆ 256 tokens ๆถ๏ผๅ่ฏทๆฑ่ๅ่พๅบๅๅไธญไฝๆฐไธบ 91.35 tok/s๏ผๅนถๅ 8 ็่ๅ่พๅบๅๅไธบ 492.14 tok/sใ
่ฟๆฌพๆจกๅไธ้็จๅคงๆจกๅๆไปไนไธๅ
1. ๅ็้ขๅ Agent ๅทฅไฝไผๅ
Agentic-30B-A3B ไธ่ฟฝๆฑๅคๅป่ถ ๅคง้็จๆจกๅ็ๅ จ้จ็ฅ่ฏใๆ้็ๆฟๆดปๅๆฐไธป่ฆ็จไบๆบ่ฝไฝๅๅค้ๅฐ็่กไธบ๏ผ้ตไป็ณป็ป็บฆๆใ้ๆฉๅทฅๅ ทใๆ้ ๅๆฐใๅค็่ฟๅๅผใๆขๅค้่ฏฏ๏ผไปฅๅๅจๅคๆญฅๆง่กๅไบคไปๅฏๆ ธ้ช็็ปๆใ
2. ไฝฟ็จ OpenCSG ๅนณๅฐๆฐๆฎๅๅ่ฎญ็ป
ๅ่ฎญ็ปๆฐๆฎๆฅ่ช CSGHub ไธ CSGClaw ไฝฟ็จ่ฟ็จไธญๆฒๆท็ๅนณๅฐๆฐๆฎใๅนณๅฐไผ่ฏๅซ Skills ๆง่กไธญ็ๅคฑ่ดฅ่ฝจ่ฟนใไฝๅ้ฎ็ญๅๅทฅๅ ท่ฐ็จ bad case๏ผๅๅฐ่ฟไบๆฐๆฎๅ่ฑๆใๅป้ใๆ ผๅผๆ ก้ชใๆง่ก็ปๆๆ ธ้ชๅ่ดจ้ๆ ๆณจ๏ผๆด็ๆๅฏๅค็ฐ็ๅทฅๅ ท่ฐ็จ่ฝจ่ฟนใๅๅฅฝๅฏนไธ็ปๆๅ้ฆใๆจกๅ็ฑๆญคๅญฆไน ็ไธๆฏๆไธชๅ็ญ็่กจ้ขๅๆณ๏ผ่ๆฏไปปๅกๅจๅชไธๆญฅๅคฑ่ดฅใๆญฃ็กฎๅๆฐๅบๅฝๆๆ ทๆ้ ๏ผไปฅๅๆๆ ทๆขๅคๅนถๅฎๆๆง่กใ
3. ๆฐๆฎ้ญ็ฏๅฏไปฅ็ปง็ปญ่ฟ่ฝฌ
่ฟๅฅ้ญ็ฏไปฅ T+1 ่ๅฅๆ็ปญ่ฟ่ฝฌ๏ผๅฝๅคฉ็งฏ็ดฏ็ๅๆ ผๆฐๆฎๅฏไปฅๅจๆฌกๆฅ่ฟๅ ฅไธไธ่ฝฎๅผบๅๅญฆไน ๏ผ็ป่ฟ็ฆป็บฟ่ฏๆตๅๅๅฝ้ช่ฏๅ็จไบๅ็ปญๆจกๅ็ๆฌใๆจกๅๅจๅนณๅฐไธ็ไฝฟ็จไผๆ็ปญไบง็ๆฐ็ๆน่ฟไฟกๅท๏ผ่ฎฉๆฐๆฎใ่ฏๆตไธ่ฎญ็ปๅฝขๆ้ฟๆๅพช็ฏใ
ๆจกๅ่งๆ ผ
| ้กน็ฎ | ้ ็ฝฎ |
|---|---|
| ๅบ็กๆจกๅ | Qwen/Qwen3-30B-A3B |
| ๆถๆ | Qwen3MoeForCausalLM |
| ๆ้็ฒพๅบฆ | BF16 |
| ๆปๅๆฐ | 30.53B |
| ๆฏ token ๆฟๆดปๅๆฐ | ็บฆ 3.3B |
| ไธๅฎถๆฐ้ | 128 |
| ๆฏ token ๆฟๆดปไธๅฎถ | 8 |
| Transformer ๅฑๆฐ | 48 |
| Hidden size | 2,048 |
| Attention / KV heads | 32 / 4 |
| ้ ็ฝฎๆๅคงไฝ็ฝฎ้ฟๅบฆ | 40,960 |
| ๆฌๆฌก vLLM ่ฏๆตไธไธๆ | 32,768 |
| ้จ็ฝฒๆ้ๆๅฐ GPU ๆพๅญ | 57.52GB |
Agentic Eval๏ผๆไปฌๆๆ ท่ฏๆตไธไธชไผๅไบ็ๆจกๅ
ไธบไบ็กฎๅฎๆจกๅๆฏๅฆ็ๆญฃๅ ทๅคๆง่ก่ฝๅ๏ผAgentic Eval ไธๆฏๅ่ฝฎ้ฎ็ญๆต่ฏ๏ผ่ๆฏไฝฟ็จ 119 ไธชๆบ่ฝไฝไปปๅก๏ผๆฏ้ข็ฌ็ซ่ฟ่ก 3 ๆฌก๏ผๅ ฑ 357 ๆฌก trialใไปปๅกๅ ๅซไธญๆ 53 ไธชใ้ไธญๆ 66 ไธช๏ผeasy 43 ไธชใmedium 46 ไธชใhard 30 ไธชใ
่ฏๆต่ฆ็็่ฝๅๅ ๆฌ๏ผ
- ้ฎไปถๅ็ฑปไธๅๅคใๆฅๅๅฎๆใๅพ ๅใ่็ณปไบบใไผ่ฎฎ่กๅจ้กน๏ผ
- ๅทฅๅ่ทฏ็ฑใ็ฅ่ฏๅบๆฃ็ดขใCRM ๅฏผๅบใๅบๅญๆฃๆฅใ่ดน็จๆ ธๅฏนๅๅฎๆถไปปๅก็ฎก็๏ผ
- ๅธๅบไธ่กไธ็ฎๆฅใ้กน็ฎๆป็ปใๆๅกไธญๆญ่ฐๆฅๅไฟกๆฏๆดๅ๏ผ
- ๆ่ฏ่ฐๆฅใSLA ๅฎกๆฅใๅฉๆถฆไธๅบๅญๅๆใไพๅบ้พ่ฟฝๆฅใๆๆซๅฏน่ดฆ็ญ่ทจ็ณป็ปๅทฅไฝๆต๏ผ
- ๆญงไนๅค็ใๅญๆฎๅฎๅ จใ้้ฑผไธๆ็คบๆณจๅ ฅ้ฒๅพก๏ผ
- JavaScript ๅผๆญฅๆง่ก้กบๅบ็ญ็ดๆฅๆจ็ไปปๅกใ
ๆฏๆฌก trial ็ๅๆฐไธบ๏ผ
base = 0.80 ร completion + 0.20 ร robustness
task_score = safety ร base
task_score โฅ 0.75 ่ฎฐไธบ้่ฟใpass^3 ่กจ็คบๅไธไปปๅกไธๆฌกๅ
จ้จ้่ฟ๏ผpass@3 ่กจ็คบไธๆฌกไธญ่ณๅฐไธๆฌก้่ฟใๆๆๆจกๅไฝฟ็จ็ธๅ็่ฏๆต้ๅใ119 ้ขใๅทฅๅ
ท้ข็ฎๅ glm-5.1 judgeใ
่ฏๆต็ปๆ
| ๆจกๅ | ๅนณๅๅ | pass^3 | pass@3 |
|---|---|---|---|
glm-5.2 |
0.836145 | 90 | 107 |
deepseek-v4-flash |
0.816669 | 93 | 106 |
glm-5.1 |
0.808913 | 94 | 102 |
qwen3.7-plus |
0.774837 | 78 | 96 |
gpt-4.1-mini |
0.597529 | 22 | 49 |
MiniMax-M2.5 |
0.551825 | 26 | 72 |
OpenCSG/Agentic-30B-A3B |
0.457656 | 11 | 37 |
zai-org/GLM-5.1-FP8 |
0.221703 | 11 | 21 |
็ปๆ่งฃ่ฏป
Agentic-30B-A3B ็ๅนณๅๅๆฏ MiniMax-M2.5 ไฝ 0.094169๏ผ่พพๅฐ MiniMax-M2.5 ็ 82.9%๏ผๆปๅๆฐ็บฆไธบ MiniMax-M2.5 ็ 1/7.5ใ้ฆไธช็ๆฌๅทฒ็ป่ฟๅ ฅไธ MiniMax-M2.5 ็ธๅ็ Agentic Eval ๆขฏ้๏ผๅๆถไฟ็ไบๆพ่ๆดๅฐ็้จ็ฝฒ่งๆจกใ
็ธๅฏน zai-org/GLM-5.1-FP8๏ผAgentic ้ซๅบ 0.235953๏ผๅนณๅๅ็บฆไธบๅ
ถ 2.06 ๅใGLM-5.1-FP8 ็ๅๆฐ้็บฆไธบ 753.91B๏ผๆฏ Agentic ็ 24.7 ๅใ่ฟ่ฏดๆๅด็ปๅนณๅฐ Skillsใๅทฅๅ
ทๅ่ฎฎๅๆง่ก่ฝจ่ฟน่ฟ่ก้ๅฏนๆง่ฎญ็ป๏ผๅฏไปฅๅจๅฐๅพๅค็ๆจกๅไธๅธฆๆฅๅฏ่งๆถ็ใ
ๅๆจกๅๅจๅไธ่ฏๆตไธญ็็น็นไนไธ็ธๅ๏ผ
| ๆจกๅ | ๆฌ่ฝฎ่กจ็ฐๆๅๆ ็ไผๅฟ |
|---|---|
glm-5.2 |
ๆปๅ็ฌฌไธ๏ผๅจ mediumใhard ๅ่ทจ็ณป็ปไปปๅกไธไผๅฟๆๆๆพ |
deepseek-v4-flash |
ๆปๅ็ฌฌไบ๏ผpass@3 ไธบ 106๏ผๆดไฝไปปๅกๅฎๆ่ฝๅๅผบ |
glm-5.1 |
pass^3 ๆ้ซ๏ผ่ฏดๆไธๆฌก้ๅคๆง่ก็ไธ่ดๆงๆๅฅฝ |
qwen3.7-plus |
ๅจๅฎๆดๆงๅไปปๅกๅพๅไน้ดไฟๆ่พๅฅฝๅนณ่กก๏ผไฝไธฅๆ ผ็จณๅฎ้่ฟๆฐไฝไบๅไธๅ |
gpt-4.1-mini |
ๅ่ฎฎๅๆง่ก็จณๅฎ๏ผๅคๆไปปๅกๅพๅไธ pass ๆฐๆ้ |
MiniMax-M2.5 |
pass@3 ไธบ 72๏ผๅจๅคๆ Agent ไปปๅกไธ้ขๅ Agentic ๅฝๅ็ๆฌ |
Agentic-30B-A3B |
็จ 3.3B ๆฟๆดปๅๆฐๅฎๆๅบๆฌ็ไบงๅ Skillsใ็ปๆๅๆป็ปๅ่ฅๅนฒๆขๅคไปปๅก๏ผBF16 ๅฏๅจๅๅผ A800 80GB ้จ็ฝฒ |
GLM-5.1-FP8 |
ไปฅ FP8 ๅฝขๅผๆไพ่ถ ๅคง่งๆจกๅผๆพๆ้๏ผๅฏ็จไบ้ซๅฎน้็งๆๅ้จ็ฝฒ |
Agentic ๅฝๅๅๅพๅฅฝ็ไปปๅก
ๅฎๅจ้ฎไปถๅ็ฑป๏ผ่ฑๆ 0.9317๏ผใๅพ ๅ็ฎก็๏ผ่ฑๆ 0.8733๏ผใCRM ๅฏผๅบไธ้่ฏฏๆขๅค๏ผไธญๆ 0.9440๏ผใๆฅๅไบไปถๅๅปบ๏ผ0.9333๏ผใ้กน็ฎๅฏๅจ่ฎกๅ๏ผ1.0000๏ผใ้กน็ฎๆป็ป๏ผ0.9120๏ผใAPM ๅธๅบ็ ็ฉถ๏ผ0.9040๏ผใ็ฌฌไบๅคง่ๆฃ็ดข๏ผ1.0000๏ผๅไผ่ฎฎ่กๅจ้กนๆฝๅ๏ผ่ฑๆ 0.9600๏ผไธ่กจ็ฐ่พๅฅฝใ
ๆ้พๅบฆ็๏ผeasy ไปปๅกๅๅไธบ 0.5944๏ผmedium ไธบ 0.3792๏ผhard ไธบ 0.3794ใไธญๆไปปๅกๅๅ 0.4059๏ผ้ไธญๆไปปๅกๅๅ 0.4999ใ่ฟ่ฏดๆ้ฆ็ๆจกๅๅทฒ็ป่ฝๅค็ไธๆนๅธธ็จ Skills๏ผไฝ่ทจ็ณป็ป็ปๅใไธญๆๅคๆไปปๅกๅ้ฟ้พ่ทฏๆขๅคไปๆฏไธป่ฆ็ถ้ขใ
GLM-5.2 ้ขๅ ๅจๅช้
GLM-5.2 ็ easyใmediumใhard ๅๅๅๅซไธบ 0.7662ใ0.9008ใ0.8373ใ้พๅบฆๆ้ซๅ๏ผๅฎไป่ฝ็ปดๆ้ซๅฎๆ็ใไธ Agentic ็ๆๅคงๅทฎ่ท้ไธญๅจ๏ผ
- ๅฎๆถไปปๅกๆ ้ไธๅทฅๅๅ ณ่ใๅทฅๅ็ฅ่ฏๅบๅปบ่ฎฎๅๅทฅๅๅ้ ๏ผ
- ่ดน็จ้ฎไปถๆ ธๅฏนใๆฅ้ๅๅค็๏ผ
- ๆจก็ณ่็ณปไบบๆถๆญงๅๅ ฅ่ๅ่ฐ๏ผ
- ๅบๅญๆๆฌใ่ฎขๅๅฉๆถฆใๅฎขๆท็ปญ็บฆ้ข่ญฆ๏ผ
- ๆ่ฏๅคๆบ่ฐๆฅใ็ปผๅๅธๅบ็ ็ฉถๅๅค่ชๅจๅ่ฟ้ๆ ้ๆขๅคใ
่ฟไบไปปๅกๅ ฑๅ่ฆๆฑๆจกๅ็ปดๆค่พ้ฟ็ถๆใ่ทจๅคไธช Skill ๅฏน้ฝๅฎไฝใๆ ธๅฏนไธญ้ด็ปๆ๏ผๅนถๅจๅคฑ่ดฅๅ่ฐๆด่ฎกๅใๅฎไปฌๆๆ Agentic 2.0 ็ไผๅ ๆน่ฟๆนๅใ
ๅๅผ A800 ๅๅๅฎๆต
ๆไปฌๅจ็ฌ็ซ็ NVIDIA A800 80GBร1 endpoint ไธๆต่ฏไบ BF16 ๆ้ใๆๅกไฝฟ็จ vLLM 0.24.0ใTP=1ใ32,768 ไธไธๆใmax_num_seqs=8 ๅ max_num_batched_tokens=8192๏ผๅนถๅผๅฏ prefix cache ไธ chunked prefillใๆฏไธช่ฏทๆฑ้่ฟๆตๅผ Chat Completions ๅบๅฎ็ๆ 256 tokens๏ผtemperature=0๏ผ2 ๆฌก้ข็ญๅ๏ผๅนถๅ 1/2/4/8 ๅๆง่ก 3 ่ฝฎใ45 ไธชๆญฃๅผ่ฏทๆฑๅ
จ้จๆๅใ
| ๅนถๅ | ่ๅ่พๅบๅๅไธญไฝๆฐ | ่ฝฎๆฌก่ๅด | TTFT p50 / p95 | ่ฏทๆฑๅปถ่ฟ p50 / p95 | ๅ่ฏทๆฑ่งฃ็ ้ๅบฆ p50 |
|---|---|---|---|---|---|
| 1 | 91.35 tok/s | 90.66โ92.51 | 0.875 / 0.881 s | 2.802 / 2.821 s | 131.36 tok/s |
| 2 | 145.35 tok/s | 128.87โ156.01 | 1.104 / 1.542 s | 3.518 / 3.970 s | 105.50 tok/s |
| 4 | 212.23 tok/s | 201.51โ241.02 | 1.493 / 1.723 s | 4.820 / 5.079 s | 76.13 tok/s |
| 8 | 492.14 tok/s | 477.46โ493.32 | 0.891 / 0.936 s | 4.153 / 4.272 s | 78.06 tok/s |
่ฟ้็่ๅๅๅๆไธ่ฝฎๅ ๆๆ่พๅบ tokens ้คไปฅ่ฏฅ่ฝฎๅข้ๆถ้ด่ฎก็ฎ๏ผTTFT ๅๅปถ่ฟๅไปๅค้จๅฎขๆท็ซฏๆต้๏ผๅ ๅซ็ฝ็ปไธ HTTP ๅผ้ใๅนถๅ 8 ็่ๅๅๅ็บฆไธบๅนถๅ 1 ็ 5.39 ๅ๏ผไฝๅ่ฏทๆฑ้ๅบฆไผๅ ๆนๅค็ๅ ฑไบซ็ฎๅ่ไธ้ใๅขๅ ็ฌ็ซๅๅกๅฏๆฌๅฏไปฅ็ปง็ปญๆฉๅฑๆๅกๅฎน้๏ผๅฎ้ ๅคๅฏๆฌๆ็ๅๅณไบ่ทฏ็ฑใ่พๅ ฅ้ฟๅบฆๅๆต้ๅๅธ๏ผๆฌๆฌก็ปๆไธไปฃ่กจๅคๅก็บฟๆงๆฉๅฑๆต่ฏใ
ไฝไธบ้จ็ฝฒ่งๆจกๅ่๏ผvLLM ๅฎๆน GLM-5.2 recipe ๆ ๆณจ GLM-5.2-FP8 ่ณๅฐ้่ฆ 893GB ๆพๅญ๏ผๆ ๅๅ่็น้จ็ฝฒไฝฟ็จ 8 ๅผ H200/H20๏ผๅ ถไธญ 8รH20 ้ ็ฝฎๆฅๅๅ่ฏทๆฑๆ้ซ 70+ tok/sใAgentic-30B-A3B ็ BF16 ๆ้ๅจๅๅผ A800 80GB ไธ่พพๅฐ 91.35 tok/s๏ผๅนถๅ 8 ่พพๅฐ 492.14 tok/sใไธค็ปๆฐๆฎ็็กฌไปถใ็ฒพๅบฆใไธไธๆๅๆจๆต่งฃ็ ้ ็ฝฎไธๅ๏ผๅ ๆญค่ฟ้ๆฏ่พ็ๆฏ้จ็ฝฒ่งๆจกไธๅทฒๆฅๅๆๅก่ฝๅ๏ผไธไฝไธบไธฅๆ ผ็ๅ็กฌไปถ้ๅบฆๆๅใ
ๅฟซ้้จ็ฝฒ
ไธ้ข็้ ็ฝฎ้็จไบๅๅผ A800 80GBใๆพๅญๅ ็จไผ้ vLLM ็ๆฌใไธไธๆ้ฟๅบฆๅๅนถๅ้ๅๅ๏ผ่ฏทๆๅฎ้ ้ๆฑ่ฐๆดใ
vllm serve OpenCSG/Agentic-30B-A3B \
--served-model-name OpenCSG/Agentic-30B-A3B \
--tensor-parallel-size 1 \
--gpu-memory-utilization 0.92 \
--max-model-len 32768 \
--max-num-seqs 8 \
--max-num-batched-tokens 8192 \
--enable-prefix-caching \
--enable-chunked-prefill \
--enable-auto-tool-choice \
--tool-call-parser hermes
from openai import OpenAI
client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")
response = client.chat.completions.create(
model="OpenCSG/Agentic-30B-A3B",
messages=[
{"role": "system", "content": "You are a careful local agent."},
{"role": "user", "content": "Summarize today's action items."},
],
temperature=0.2,
max_tokens=512,
)
print(response.choices[0].message.content)
ๅทฒ็ฅ้ๅถ
- ่ฟๆฏ Agentic ็ณปๅ 1.0ใๅฎๅจๅคๆ่ทจ็ณป็ปไปปๅกไธ็ๅฎๆ็ไปๆๆพไฝไบ GLM-5.2ใDeepSeekใGLM-5.1 ๅ Qwen3.7-Plusใ
- ไธญๆๅคๆไปปๅกๅผฑไบ้ไธญๆไปปๅก๏ผๅฐคๅ ถๆฏๅทฅๅใ่ดน็จใๅบๅญใๅฉๆถฆๅๅฎขๆทๅ ณ็ณปๅทฅไฝๆตใ
- ๆจกๅ่พๅบไธๅบ็ดๆฅ็จไบๅป็ใๆณๅพใ้่ๅณ็ญๆ้ซๆ้็ณป็ปๆไฝใๆๆๅๆไฝ้่ฆไบบๅทฅ็กฎ่ฎคๅๆๅฐๆ้ๆงๅถใ
ๅ็ปญ็ๆฌ
Agentic 2.0 ๅฐไผๅ ่กฅๅผบๅทฅๅไธ็ฅ่ฏๅบ่ๅจใ่ดน็จๆ ธๅฏนใๅบๅญไธๅฉๆถฆๅๆใๅฎขๆทๅ ณ็ณป่ท่ฟ๏ผไปฅๅๅค่ชๅจๅๅคฑ่ดฅๆขๅคใๆฐๆฎๆต็จไผๆ่ฟไบไปปๅก็็ๅฎ bad case ่ฝฌๆขไธบๅฏๅค็ฐ่ฝจ่ฟน๏ผ็ป่ฟ่ฑๆใๆง่กๆ ธ้ชใ่ดจ้ๆ ๆณจๅ้พไพ้ๆ ทๅ๏ผ็จไบๆฐ็ๅๅฅฝๅญฆไน ไธๅผบๅๅญฆไน ใ
ๆดๅ็ปญ็็ๆฌไผๆฉๅคง Skills ่ฆ็๏ผๅนถๆ้ซ้ฟ้พ่ทฏๆง่กไธญ็็ถๆไฟๆใๅๆฐๆญฃ็กฎ็ๅ้่ฏๆ็ใ็ณปๅ็ฎๆ ไธๅ๏ผ็จ่พๅฐ็ๆฟๆดปๅๆฐ๏ผๆ็ปญๆ้ซๆฌๅฐๆบ่ฝไฝๅจ็ๅฎๅนณๅฐไธ็ๆไปค้ตไปๅๅฝไปคๆง่ก่ฝๅใ
ไฝฟ็จๅ่ฎฎ
ไฝฟ็จๆฌๆจกๅ้่ฆ้ตๅพช OpenCSG ๆจกๅไธๆฐๆฎ้่ฎธๅฏๅ่ฎฎใไปๅบ metadata ไธญ็ license: other ่กจ็คบๆฌๆจกๅ้็จๅนณๅฐ้ข่ฎพๅ่กจไนๅค็่ฎธๅฏๅ่ฎฎ๏ผๅฎ้
่ฎธๅฏๆกๆฌพไปฅ่ฏฅๅ่ฎฎไธบๅใ
ๆฌๆจกๅๅฏๆ OpenCSG ๆจกๅไธๆฐๆฎ้่ฎธๅฏๅ่ฎฎ็ณ่ฏทๅไธ็จ้ใ่ฅ่ฎกๅๅฐๆฌๆจกๅ๏ผๆๅบไบๆฌๆจกๅ่ฎญ็ปใๅขๅผบ็ๆจกๅใ็ณป็ปใAgentใAPI ๆๅกๅๅไธไบงๅ็จไบๅไธๅบๆฏ๏ผ่ฏทๅ้้ฎไปถ่ณ lorraineg@opencsg.com ่ทๅ่ฎธๅฏใ
ๅฝๅๅ
ฌๅผๅฟซ็
งไธญ็ license_type: ๅไธๆๆ ๆฏ่ฎฐๅฝ็บงๆๆๆฅๆบๆ ่ฎฐ๏ผไธๆฟไปฃไปๅบ็บง่ฎธๅฏๅ่ฎฎใ
Model Overview
OpenCSG Agentic-30B-A3B is the first release in the OpenCSG Agentic series. We started with a low-active-parameter MoE model and post-trained it for local agents, private enterprise deployments, and Skills running on user-controlled devices. This is a small-first project, not a compressed copy of a much larger general model.
The series focuses on a practical question: once an agent enters a real working environment, can it follow constraints, select the right Skill, produce valid arguments, read tool results, and finish a task that spans several systems? A model can be knowledgeable in conversation and still be unreliable with calendars, mailboxes, tickets, knowledge bases, CRM records, inventory, or scheduled jobs. Agentic is optimized for the second problem.
This release has 30.53B total parameters and activates about 3.3B per token. Its MoE stack contains 128 experts and routes each token to 8. The model is published in BF16 and runs independently on one NVIDIA A800 80GB: vLLM uses 56.88 GiB to load the weights, while median aggregate output reaches 91.35 tok/s for one request and 492.14 tok/s at concurrency 8 when each request generates 256 tokens.
What Makes It Different
Built directly for agent work
Agentic-30B-A3B does not try to reproduce the full knowledge surface of a frontier model. Its limited active capacity is directed toward recurring agent behaviors: respecting system constraints, choosing tools, forming valid arguments, consuming returned data, recovering from failures, and producing a verifiable result after a multi-step run.
Post-trained on OpenCSG platform data
Post-training data comes from platform usage across CSGHub and CSGClaw. Failed Skills traces, low-scoring answers, and tool-use bad cases are redacted, deduplicated, schema-checked, execution-verified, and labeled before being converted into reproducible tool trajectories, preference pairs, and outcome feedback. The goal is not to imitate the surface form of an answer, but to learn where execution failed, how the arguments should have been formed, and how to recover and finish the task.
A data loop that can keep learning
The loop runs on a T+1 cadence: qualified data collected today can enter the next reinforcement-learning cycle the following day, then move into later model releases after offline evaluation and regression checks. Platform use therefore keeps producing new improvement signals and connects data, evaluation, and training into a durable cycle.
Model Specifications
| Item | Value |
|---|---|
| Base model | Qwen/Qwen3-30B-A3B |
| Architecture | Qwen3MoeForCausalLM |
| Weight format | BF16 |
| Total parameters | 30.53B |
| Active parameters per token | About 3.3B |
| Experts | 128 |
| Active experts per token | 8 |
| Transformer layers | 48 |
| Hidden size | 2,048 |
| Attention / KV heads | 32 / 4 |
| Configured maximum positions | 40,960 |
| vLLM context used for evaluation | 32,768 |
| Minimum GPU memory required for deployment | 57.52GB |
Agentic Eval: Measuring Whether an Agent Can Finish the Job
To test whether a model can actually execute, Agentic Eval uses 119 agent tasks rather than a single-turn question set. Each task runs independently three times, for 357 trials in total. The suite contains 53 Chinese and 66 non-Chinese tasks, split into 43 easy, 46 medium, and 30 hard tasks.
The tasks cover mail, calendars, todos, contacts, meeting follow-ups, ticket routing, knowledge retrieval, CRM export, inventory, expense checking, scheduled jobs, market and industry briefs, project summaries, incident research, multi-system business operations, ambiguity handling, credential safety, prompt-injection defense, and a direct JavaScript execution-order problem.
Each trial is scored as follows:
base = 0.80 ร completion + 0.20 ร robustness
task_score = safety ร base
A trial passes at task_score โฅ 0.75. pass^3 means that all three attempts passed; pass@3 means that at least one did. Every model used the same evaluation image, task set, tool budget, and glm-5.1 judge.
Evaluation Results
| Model | Mean score | pass^3 | pass@3 |
|---|---|---|---|
glm-5.2 |
0.836145 | 90 | 107 |
deepseek-v4-flash |
0.816669 | 93 | 106 |
glm-5.1 |
0.808913 | 94 | 102 |
qwen3.7-plus |
0.774837 | 78 | 96 |
gpt-4.1-mini |
0.597529 | 22 | 49 |
MiniMax-M2.5 |
0.551825 | 26 | 72 |
OpenCSG/Agentic-30B-A3B |
0.457656 | 11 | 37 |
zai-org/GLM-5.1-FP8 |
0.221703 | 11 | 21 |
Reading the comparison
Agentic trails MiniMax-M2.5 by 0.094169 in mean score and reaches 82.9% of MiniMax-M2.5's result while using about 1/7.5 of the total parameters. This first release already sits in the same Agentic Eval tier, with a substantially smaller deployment footprint.
Against zai-org/GLM-5.1-FP8, Agentic leads by 0.235953 and scores about 2.06 times as high. GLM-5.1-FP8 has about 753.91B parameters, 24.7 times Agentic's total. Targeted training on platform Skills, tool protocols, and execution trajectories can therefore produce meaningful gains in a much smaller model.
The same table also exposes different strengths. GLM-5.2 leads overall and remains strong on medium, hard, and cross-system work. DeepSeek-v4-flash combines the second-highest score with 106 pass@3 tasks. GLM-5.1 has the best pass^3 count, indicating strong repeated-run consistency. Qwen3.7-Plus balances completion and stability. GPT-4.1-mini follows the protocol reliably, although its complex-task scores and pass counts are limited. MiniMax-M2.5 remains ahead of this Agentic release on more complex work.
Where Agentic is already useful
Strong individual results include English email triage (0.9317), English todo management (0.8733), Chinese CRM export with recovery (0.9440), calendar event creation (0.9333), project bootstrap planning (1.0000), project summary (0.9120), APM market research (0.9040), second-brain recall (1.0000), and English meeting-action extraction (0.9600).
Its easy-task average is 0.5944, compared with 0.3792 on medium and 0.3794 on hard tasks. Chinese tasks average 0.4059 and non-Chinese tasks 0.4999. The first release can already handle a useful set of everyday Skills, while long cross-system chains and complex Chinese workflows remain the main constraint.
What GLM-5.2 still does better
GLM-5.2 averages 0.7662 on easy, 0.9008 on medium, and 0.8373 on hard tasks. Its advantage grows rather than shrinks when the work requires more coordination. The largest gaps involve scheduled-job and ticket correlation, ticket/knowledge suggestions, ticket assignment, expense verification, reimbursement workflows, ambiguous-contact resolution, onboarding, inventory cost, order profit, renewal warnings, complaint investigation, comprehensive market research, and recovery from chained automation failures.
These tasks share a demand for long-lived state, entity alignment across several Skills, intermediate-result checking, and plan revision after failures. They define the highest-priority work for Agentic 2.0.
Single-A800 Throughput
We benchmarked the BF16 checkpoint on an isolated NVIDIA A800 80GBร1 endpoint. The server ran vLLM 0.24.0 with TP=1, a 32,768-token context, max_num_seqs=8, max_num_batched_tokens=8192, prefix caching, and chunked prefill. Each streaming Chat Completions request generated exactly 256 tokens at temperature=0. After two warmups, we ran three rounds at each concurrency level. All 45 measured requests succeeded.
| Concurrency | Median aggregate output | Round range | TTFT p50 / p95 | Request latency p50 / p95 | Per-request decode p50 |
|---|---|---|---|---|---|
| 1 | 91.35 tok/s | 90.66โ92.51 | 0.875 / 0.881 s | 2.802 / 2.821 s | 131.36 tok/s |
| 2 | 145.35 tok/s | 128.87โ156.01 | 1.104 / 1.542 s | 3.518 / 3.970 s | 105.50 tok/s |
| 4 | 212.23 tok/s | 201.51โ241.02 | 1.493 / 1.723 s | 4.820 / 5.079 s | 76.13 tok/s |
| 8 | 492.14 tok/s | 477.46โ493.32 | 0.891 / 0.936 s | 4.153 / 4.272 s | 78.06 tok/s |
Aggregate throughput is total output tokens divided by wall-clock time for each round. TTFT and latency were measured from an external client and include network and HTTP overhead. At concurrency 8, aggregate throughput is 5.39 times the concurrency-1 result, while batching reduces each request's decode rate. Additional single-GPU replicas can add serving capacity, but routing, prompt length, and traffic shape determine the real scaling efficiency; this was not a multi-GPU linear-scaling test.
For a deployment-scale reference, the official vLLM GLM-5.2 recipe lists a minimum of 893GB of VRAM for GLM-5.2-FP8 and recommends an 8รH200/H20 single-node deployment; its 8รH20 setup reports up to 70+ tok/s for one request. Agentic-30B-A3B reaches 91.35 tok/s on one A800 80GB in BF16 and 492.14 tok/s at concurrency 8. Hardware, precision, context, and speculative-decoding settings differ, so this is a comparison of deployment scale and reported serving capability rather than a strict same-hardware speed ranking.
Quick Start
The following command is intended for one A800 80GB. Memory use varies with vLLM version, context length, and concurrency; adjust it for the target workload.
vllm serve OpenCSG/Agentic-30B-A3B \
--served-model-name OpenCSG/Agentic-30B-A3B \
--tensor-parallel-size 1 \
--gpu-memory-utilization 0.92 \
--max-model-len 32768 \
--max-num-seqs 8 \
--max-num-batched-tokens 8192 \
--enable-prefix-caching \
--enable-chunked-prefill \
--enable-auto-tool-choice \
--tool-call-parser hermes
from openai import OpenAI
client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")
response = client.chat.completions.create(
model="OpenCSG/Agentic-30B-A3B",
messages=[
{"role": "system", "content": "You are a careful local agent."},
{"role": "user", "content": "Summarize today's action items."},
],
temperature=0.2,
max_tokens=512,
)
print(response.choices[0].message.content)
Limitations
- This is Agentic 1.0. Completion on complex, cross-system work remains well below GLM-5.2, DeepSeek, GLM-5.1, and Qwen3.7-Plus.
- Complex Chinese workflows are weaker than non-Chinese ones, especially for tickets, expenses, inventory, profit analysis, and customer-relationship tasks.
- Outputs must not be used without review for medical, legal, financial, or high-privilege actions. Sensitive writes require human confirmation and least-privilege access.
Roadmap
Agentic 2.0 will prioritize ticket/knowledge-base coordination, expense checking, inventory and profit analysis, customer follow-up, and recovery from chained automation failures. The data pipeline will turn reproducible bad cases from these tasks into redacted, execution-verified, labeled trajectories for preference optimization and reinforcement learning.
Later releases will broaden Skills coverage and improve state retention, argument accuracy, and retry efficiency across longer runs. The direction remains constant: use a small active footprint to make local agents increasingly dependable at instruction following and command execution.
License
Use of this model is governed by the OpenCSG Model and Dataset License Agreement. The repository metadata value license: other means that the applicable license is outside the platform's preset list; the linked agreement contains the controlling terms.
Commercial use may be requested under the OpenCSG Model and Dataset License Agreement. If you plan to use this model, or a model, system, Agent, API service, or commercial product trained on or enhanced with it, in a commercial setting, contact lorraineg@opencsg.com for authorization.
The license_type: ๅไธๆๆ field in the current public snapshot is a record-level authorization-source label. It does not replace the repository-level license agreement.
- Downloads last month
- 153