Ganymede A1

Kozu AI Research · 27B class · Apache-2.0BF16 weightsReasoning + instructionReport GA1-2026.07More correct answers. Less reasoning overhead.

Release brief

01

Ganymede A1 is the high-capacity member of the Kozu reasoning family. In matched open-harness evaluation, it improves five of seven reported capability rows. The headline GSM8K result reaches 0.913 with 4.7× fewer thinking tokens.

The family principle is intelligence per token: spend reasoning only where it improves the answer. Ganymede A1 is the 27B-class choice for higher-capacity work; Deimos R1 is the compact option and Europa B1 the balanced option. These are deployment roles, not cross-model benchmark claims.

4.7×fewer GSM8K think tokens+0.449MMLU-Pro absolute delta6.6×held-out reasoning gain · n=200

Evaluation

02

Each row compares matched runs in the same open harness. The score table is the source of truth; the token column makes the cost of each result visible. Run-local results should not be compared with vendor-published scores from different evaluation stacks.

lm-eval 0.4.12 · chat template · few-shot as multi-turn · temperature 1.0 · top_p 0.95 · presence penalty 1.5 · thinking enabled · seed 42 · GSM8K/IFEval n=150 · MMLU-Pro n=25 per subtask.

• GSM8K and IFEval used a 16,384-token cap for Ganymede A1 and a 32,768-token cap for the reference; the cap can only constrain Ganymede A1. MMLU-Pro used a symmetric 32,768-token cap.
• The MMLU-Pro reference run produced a low score across two independent runs (0.114 / 0.054) and short mean thinking. This is a harness interaction, not a claim about the reference model’s general knowledge.
• The held-out, contamination-filtered math audit with verified ground truth moved from 0.065 to 0.430 at n=200. The 12-check instruction probe moved from 1.000 to 0.917 and is directional only.
• Strict GSM8K rewards silently copying the few-shot #### N ending. Explicitly requested formats are more reliable; request the format in words.
• Raw evaluation artifacts are retained under bench/.

Data & attribution

03

kozu_reasoning_v1.1 is a roughly 10k-example blend of verified reasoning traces and human-authored instruction data, including a dedicated 800-example format-adherence slice.

Reasoning · ~5kKozu’s Kuiper trace inverter was designed to recreate Mythos/Fable-style reasoning from answer-scrubbed human reference solutions in GSM8K and NuminaMath 1.5. Retained traces passed symbolic answer verification, derivation checks, deterministic defect scans, and an adversarial judge panel.Instruction · ~5kHuman-authored pairs from Databricks Dolly 15k and OpenAssistant OASST2.Format · 800Deterministic rewraps of verified rows for explicitly requested formats, including JSON, boxed answers, and named answer markers.Source licensesGSM8K: MIT. NuminaMath 1.5: Apache-2.0. Databricks Dolly 15k: CC-BY-SA-3.0. OpenAssistant OASST2: Apache-2.0. Reasoning traces were generated with Kuiper-R1 by Kozu AI.

Deployment

04

The repository includes merged BF16 weights, tokenizer assets, processor configuration, generation configuration, and the chat template. Thinking appears inside <think>…</think> before the final answer.

vllm serve Michael-Kozu/Ganymede-A1 --served-model-name ganymede-a1 \
  --max-model-len 8192 --gpu-memory-utilization 0.85 --trust-remote-code

Recommended sampling: temperature 0.6–0.8, top_p 0.95, and repetition_penalty about 1.05. Healthy responses are typically well under 1k thinking tokens.

Limitations

05

• This is a research release, not a safety-certified or production-guaranteed system.
• At temperature 1.0 with presence penalty, roughly 2–3% of prompts can produce degenerate long generations toward the cap. Use the recommended sampling settings or a max-token cap of at least 8k.
• Reasoning is habitually short and does not reliably scale up for extremely hard problems. Coding and tool use are not established by this release.
• Narrow constraints such as letter avoidance succeed only about 60–80% in internal probing. Verify outputs for consequential decisions and domain-specific use.

Kozu AI · Ganymede A1Apache-2.0 · Model release