GLM-4.1V-9B-Thinking
MIT-licensed 9B reasoning vision-language model with 64K context.
About
Zhipu's first reasoning-focused vision-language model pairs the GLM-4-9B base with chain-of-thought reasoning trained via reinforcement learning with curriculum sampling. At 9B parameters with 64K context and 4K image support, it is meant to run on a single high-end GPU, and the HF card carries the MIT license.
Zhipu reports it beating the 72B Qwen2.5-VL on 18 benchmarks — plausible for a reasoning-tuned model, but vendor-reported. Zhipu's rapid version churn (4.1V, 4.5V, 4.6V) also means naming confusion when picking a checkpoint.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
- Beats Qwen2.5-VL-72B on 18 benchmarks (Zhipu-reported)
- State of the art among ~10B VLMs (Zhipu-reported)
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: MIT license per HF README metadata; released 2025-07-01 with technical report; 64K context, 4K image support, RLCS training; Zhipu's rapid version churn makes checkpoint selection confusing; Benchmarks are Zhipu-reported. License: MIT per Hugging Face README metadata on zai-org/GLM-4.1V-9B-Thinking..
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 8.0/10
- · MIT license per HF README metadata; released 2025-07-01 with technical report
- · 64K context, 4K image support, RLCS training
- · Zhipu's rapid version churn makes checkpoint selection confusing
- · Benchmarks are Zhipu-reported