GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
JamieBradfield
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
JamieBradfield
Model Tree
empero-ai/Qwen3.8-9B-Distill
Input Modalities
Output Modalities
Supported Functionality
Read this before you use the numbers below.
After this checkpoint was published, we tested the base model
(Empero/Qwen3.8-9B-Distill, no fine-tuning) through the live Hermes
interface: the OpenAI-native tools parameter path that Hermes actually
uses. The base model calls tools correctly on its own.
The harness that certified this checkpoint did not use that interface.
It baked tool schemas into the system text, never passed tools, and
required raw <tool_call> XML in the response. That prompt shape matches
neither the base model's own chat template nor the live Hermes request
path. Every gate score in the v26-to-v29 lineage measured this synthetic
interface.
The statement in this README that "the base distill contributes zero tool-calling" is incorrect. It measured the synthetic interface.
Corrected comparison (native tools path, same 4-tool battery, greedy):
| model | tier-1 fired | tier-1 name=todo | tier-2 todo-first | tier-3 fired (want 0) | tier-4 fired (want 0) |
|---|---|---|---|---|---|
| base (no fine-tune) | 18/20 | 17/20 | 3/10 | 5/10 | 0/10 |
| v28 ckpt-175 (this model) | 17/20 | 10/20 | 8/10 | 9/10 | 4/10 |
| armB2-175 (clean retrain) | 10/20 | 8/20 | 10/10 | 4/10 | 0/10 |
What the fine-tune lineage actually changed: todo-first planning on multi-step tasks (tier-2 name=todo 3/10 base to 8/10 v28 to 10/10 armB2). That is real, and it is the one capability the lineage added. The costs: restraint regressions (tier-3 over-trigger 5/10 to 9/10; tier-4 tool-drift on trivia 0/10 to 4/10) and, in armB2, over-suppression (tier-1 10/20).
Status: the fine-tune program is frozen as of 2026-09-03. The base model is the author's runtime default. This checkpoint stays published as the strongest measured demonstration of SFT-taught todo-first planning in the lineage, and as the artifact that exposed the harness error. The Evaluation section below describes behavior on the synthetic interface; do not read it as native-interface competence.
Reproduction: scripts/native_battery.py + scripts/todo_probes.jsonl
scripts/no_drift_probes.jsonl produce the corrected table (any
GGUF, CTX=65536 to match the serve config).qwen3.8-9b-hermes-fc-todo),
QLoRA adapter on top (envelope + tool tokens already baked in)hermes-fc-v28-data); probes
were kept virgin (zero overlap with generation or training)45 real ambiguous trajectories (seeded held-out split), each demanding one of a specific tool family. Same seed/limit for all models.
| model | fired | name_match | args_ok | format_exact | clarify |
|---|---|---|---|---|---|
| base Qwen3.8-9B-Distill | 0/45 | 0/45 | 0/45 | 0/45 | — |
| v2 (real traces) | 41/45 | 28/45 | 25/45 | 25/45 | 2/2 |
| v25 (trigger-heavy) | 26/45 |
v26 fires more (43/45 vs 41/45) and formats more exactly (27/45) than v2,
with clarify intact (2/2). Cost: −3 name_match / −2 args_ok — the model
sometimes picks the wrong tool under semantic ambiguity (e.g. search_files
instead of tool_describe on a docs probe). Selection precision is the
target of the next recipe iteration.
The base distill contributes zero tool-calling (0/45); the entire capability comes from the fine-tune.
Research artifact for experimenters working on tool-call behavior in 9B-class
models — not a product. The tool_call XML envelope and tool schemas match
the Hermes agent runtime this was trained on.
scripts/)All scripts that produced this model are in scripts/:
| file | purpose | number it produced |
|---|---|---|
build_v26.py | dataset build (v2-minus-lambda + SWE-rebench + APIGen + When2Call → 1,543 rows) | row count, mix |
qwen38_train_v26.py | QLoRA continue-train via Unsloth (1 epoch, 193 steps, loss 0.0552) | training loss/steps |
merge_qwen38_v26_pure.py | shard-streamed BF16 merge (pure torch; MTP head preserved; vocab patched) | 12 shards / 18.4 GB |
smoke_v25_merged.py + eval_fc_heldout.py + |
The scripts carry the author's machine paths (F:/models/..., C:/AI/...)
and run on this rig's ROCm stack (Unsloth 2026.8.18, torch 2.11 ROCm,
RX 7700 XT); adjust paths for your environment.
Convert/quantize: llama-rocmfpx fork, convert_hf_to_gguf.py --outtype bf16
→ llama-quantize Q4_0_ROCMFP4_FAST (see GGUF repo).
Q4_0_ROCMFP4_FAST GGUF (ROCmFPX format, AMD RDNA3 kernels) is published in
the companion repo. ROCmFPX quants target AMD ROCm inference; for portable
use, convert from the BF16 merge here.
The training data itself is not published. The rows derive from the author's own agent sessions, and some rows contain private strings (hostnames, session identifiers). The scripts that build the dataset are published; the data is not. The weights do not contain those strings.
Base model and license inherited from Empero/Qwen3.8-9B (Apache-2.0). Training data derived from the author's own Hermes agent sessions plus publicly-licensed teacher sources (SWE-rebench, APIGen-MT-5k, When2Call).
<|tool_call|><|tool_response|>Qwen3_5ForConditionalGeneration with
text_config.vocab_size = 248079. Visual tower dropped (text-only).| 18/45 |
| 16/45 |
| 26/45 |
| 0/2 |
| v26 (this model) | 43/45 | 25/45 | 23/45 | 27/45 | 2/2 |
convert_hf_hermes.py| 45-probe smoke / held-out eval (fired/name/args/format/clarify) |
| the A/B table above |
subagent-qwen38-workloads.py + subagent-qwen38.sh | tool-call evaluation harness (workloads A–E) | harness bench |