Results
Same harness for both models: FP8 weights, thinking on, a 64k token cap, the same prompts and sampling. Differences are paired over items, with 95% bootstrap intervals, and computed before rounding. FRAMES answers are judged by two LLMs, GLM-5.3-Flash and MiMo-V2.6-Flash; when they disagree, a rule-based check decides.
No difference is statistically significant. Document-work results are on the main model card.
Speed
Measured with vLLM 0.32.0 and 1 MTP draft token for both models, on the 20 fixed questions of the Sparkle 1.0 card (10 GPQA Diamond, 10 LiveCodeBench), one request at a time, thinking on. The median is the 11th of the 20 sorted times, as on the 1.0 card. The GPU, vLLM version and MTP setting differ from the 1.0 card's speed test, so its times are not comparable with these.
Serve with vLLM
vllm serve Egzakta/Kvark-Sparkle-1.1-27B-FP8 \ --max-model-len 131072 --kv-cache-dtype fp8 \ --reasoning-parser qwen3 \ --enable-auto-tool-choice --tool-call-parser qwen3_coder \ --speculative-config '{"method":"mtp","num_speculative_tokens":3}'
from openai import OpenAI client = OpenAI(base_url="http://localhost:8000/v1", api_key="none")response = client.chat.completions.create( model="Egzakta/Kvark-Sparkle-1.1-27B-FP8", messages=[{"role": "user", "content": "How many primes are there between 100 and 150?"}], max_tokens=65536,)message = response.choices[0].messageprint(getattr(message, "reasoning", None))print(message.content)print(response.usage.completion_tokens, "tokens")
Thinking is on by default. To turn it off for a request, pass extra_body={"chat_template_kwargs": {"enable_thinking": False}}. Sampling defaults are temperature 1.0, top_p 0.95 and top_k 20.
For document and search agents, add the stop rule (stop_system.txt in the main repository) to the end of your system prompt, after your tool instructions. The model was trained with it.
Memory
More
Training, data, document-work results and the BF16 weights are on the main model card. License: Apache 2.0, commercial use permitted, no revenue cap.