tinyanalysis

OCR models / TeleAI (XingChen)

TeleOCR

TeleAI (XingChen) · released 2026-08-17 · data checked 2026-11-06

TeleOCR is a 1.4B-parameter OCR and document-parsing model from TeleAI (XingChen) (CN), released 2026-08-17 under the Apache-2.0 license. It scores 96.91 on OmniDocBench v1.6 (#3 of 33). It needs about 3.8 GB of memory at 16-bit.

Built on Qwen2.5-VL. First released as NaviDC-OCR and renamed in September 2026.

OmniDocBench v1.6
96.91
#3 of 33 models
olmOCR-Bench
—
Not reported
Parameters
1.4B
OCR specialist
Memory to run
~3.8 GB
16-bit weights plus about 1 GB

Benchmarks

OmniDocBench v1.6 Leaderboard96.91
Best: OXR 1.0 97.03

The tick marks the best score on each benchmark. See how to read the ranking.

Specifications

Lab
TeleAI (XingChen) (CN)
Type
OCR / document parsing specialist
Total parameters
1.4B
License
Apache-2.0
Memory to run
~3.8 GB at 16-bit
Released
2026-08-17
Weights and usage on Hugging Face ↗

Closest alternatives

  • Youtu-Parsing-Omni (5.3B, 96.96 on OmniDocBench, Youtu-Parsing license)
  • Xiaomi-OCR-0 (870M, 96.83 on OmniDocBench, Apache-2.0)
  • OXR 1.0 (1.5B, 97.03 on OmniDocBench, OXR Open Source License)
  • OvisOCR3 (850M, 96.74 on OmniDocBench, Apache-2.0)

Sources

Something wrong here? Report it.