fix checking tokenizers version

[pavel-esir]

• Currently for gguf model we cannot get string representation of special tokens.
• In order to infer special tokens representation, we should run decode with skip_special_tokens=False.
• But because GGUF models don't have 'openvino_tokenizers_version' rt_info genai::Tokenizers assumes it's and old model and doesn't set states which are required in order to run decode with skip_special_tokens=False.
• We fixed getting model version, if it's a gguf then it's definitely newer than 2024.5.

Ticket: CVS-172426

[as-suvorov]

GGUF tests are failing: https://github.com/openvinotoolkit/openvino.genai/actions/runs/17265102491/job/48997729147?pr=2667#step:7:1243

[Copilot]

Pull Request Overview

This PR fixes the issue where GGUF models cannot get string representations of special tokens by properly setting version information for tokenizer models. The fix ensures that GGUF models are correctly identified as newer than version 24.5, enabling proper decode operations with special tokens.

Key changes:

• Added version information to GGUF tokenizer models to indicate they are newer than 24.5
• Fixed version checking logic to properly handle GGUF models
• Enhanced test coverage for special token handling in GGUF models

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
src/cpp/src/tokenizer/tokenizer_impl.cpp	Sets openvino_genai_version for GGUF models and fixes version checking logic
src/cpp/src/gguf_utils/gguf_tokenizer.cpp	Adds quote_meta function to properly escape special tokens in regex patterns
tests/python_tests/test_gguf_reader.py	Expands test coverage with special token prompts and token validation
.github/workflows/windows.yml	Increases timeout for GGUF reader tests from 60 to 100 minutes

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Spelling error in comment on line 128: 'utf validate' should be 'UTF validation' or 'UTF-8 validation'.