MiniCPM-V 4.6 1.3B
MiniCPM-V 4.6 1.3B is an open-weights model from OpenBMB with a 262k-token context window. No provider we track lists it yet.
Providers serving MiniCPM-V 4.6 1.3B
No provider we track lists this model yet.
MiniCPM-V 4.6 1.3B benchmark scores
Compare every model in Bench rankingsPublished scores only, each linked to its source. Ticks are every other model's best published score.
GPQA Diamond
PhD-level scientific reasoning, pass@1, no tools
30.5%, source: Artificial AnalysisJoint 209th of 221 models with a score
MMMU-Pro (standard)
Visual reasoning, 10 answer options
37.9%, source: Artificial AnalysisRanks 97th of 101 models with a score
Humanity's Last Exam (text-only)
Reasoning & knowledge, no tools
5.1%, source: Artificial AnalysisJoint 177th of 221 models with a score
τ²-Bench Telecom
Agentic tool use (telecom), with a Qwen3 model as the simulated user
87.7%, source: Artificial AnalysisRanks 44th of 160 models with a score
Terminal-Bench 2.1 (Terminus 2)
Agentic coding & terminal use
0%, source: Artificial AnalysisJoint 126th of 127 models with a score
IFBench
Instruction following, 294 prompts, 5 runs each
26.7%, source: Artificial AnalysisRanks 141st of 158 models with a score
GDPval-AA v2.1 (Elo)
Real-world work tasks; a rating against other models, not a percentage
-370, source: Artificial AnalysisRanks 141st of 144 models with a score
Terminal-Bench Hard
Agentic coding & terminal use, Artificial Analysis's 44-task hard subset
0%, source: Artificial AnalysisJoint 135th of 157 models with a score
LCR
Long-context reasoning, AA-LCR v1.1
7%, source: Artificial AnalysisJoint 183rd of 211 models with a score
Omni Accuracy
Omniscience knowledge accuracy
6.7%, source: Artificial AnalysisJoint 204th of 216 models with a score
Non-Hallucination
1 − hallucination rate
3.9%, source: Artificial AnalysisJoint 210th of 216 models with a score
Show as table
| Benchmark | Score |
|---|---|
GPQA Diamond PhD-level scientific reasoning, pass@1, no tools | 30.5%, source: Artificial Analysis |
MMMU-Pro (standard) Visual reasoning, 10 answer options | 37.9%, source: Artificial Analysis |
Humanity's Last Exam (text-only) Reasoning & knowledge, no tools | 5.1%, source: Artificial Analysis |
τ²-Bench Telecom Agentic tool use (telecom), with a Qwen3 model as the simulated user | 87.7%, source: Artificial Analysis |
Terminal-Bench 2.1 (Terminus 2) Agentic coding & terminal use | 0%, source: Artificial Analysis |
IFBench Instruction following, 294 prompts, 5 runs each | 26.7%, source: Artificial Analysis |
GDPval-AA v2.1 (Elo) Real-world work tasks; a rating against other models, not a percentage | -370, source: Artificial Analysis |
Terminal-Bench Hard Agentic coding & terminal use, Artificial Analysis's 44-task hard subset | 0%, source: Artificial Analysis |
LCR Long-context reasoning, AA-LCR v1.1 | 7%, source: Artificial Analysis |
Omni Accuracy Omniscience knowledge accuracy | 6.7%, source: Artificial Analysis |
Non-Hallucination 1 − hallucination rate | 3.9%, source: Artificial Analysis |