qwen3.5:27b
Family: qwen3.5 · 27B parameters · 1 run(s)
Fits in 24 GB VRAM · ollama 0.40.6 · Q4_K_M
Measured VRAM at 32K context (KV cache included, ollama ps on the RTX 3090): 18 GB · previous rank in this tier: #4 (results v0.1.1, before the v0.1.2 re-tiering; score unchanged). Tier rule
95.0/100 · scoring v4
95% task-bootstrap CI (main suite, 3 trials): 91.7–97.7 · pooled over 8 runs (3 main-suite trials + 5 reruns at temperature 0.7): 91.7–98.0 · ≈ statistically tied with other models in pooled tie group 2 of this tier (the leaderboard’s ≈) · rank #5 in tier, rank range under weight sensitivity #4–#5 (5th–95th percentile; min #4, max #5) · v3: 95.0 (#4) · mention 0.0% (0/45) · how
Plain-language summary
Excellent at tool-using assistant work on this suite.
Good for
- returning clean JSON
- matching the requested format
- picking the right tool
- multi-step plans
- staying safe under prompt injection
Weak at
- —
Category breakdown
| JSON | Schema | Tool | Multi-step | Injection (useful&safe) | tok/s | p50 s/task | p95 s/task |
|---|---|---|---|---|---|---|---|
| 99 | 97 | 100 | 87 | 98 | 37.8 @37becd | 22.1 | 40.3 |
Prompt injection (15 tasks × 3 trials): useful and safe 97.8% (task completed and not compromised; counts in the score) · utility under attack 97.8% (completed the user's task) · harmful-action attack success 0.0% (a real harmful tool call or canary leak). Mentions (quoting the attacker without acting): 0.0% (0/45); they do not lower the score. v3 useful and safe: 97.8%. v2.1 used injection resistance 100.0 × tool-call rate 1. Definitions
- Run:
20261022-qwen3.5-27b-d87444· 2026-10-22T11:56:01+00:00 - GPU: NVIDIA GeForce RTX 3090 (24 GB) · Linux 6.8.0-49-generic
- Host:
37becd33558a· container (RunPod) · host AMD EPYC 7402P 24-Core Processor, 251.6 GB RAM · measured 2026-10-22T11:56:01+00:00 · hosts - Harness: hermes-agent Hermes Agent v0.22.2+2363.g9cb34e9 (2026.9.24) · stable · local 9cb34e9e285d · trials 3 · sampling: Ollama default (temperature and seed not set by harness) · context window: 32768 tokens as loaded by Ollama (the unused runner field records 8192; see Methodology)
- Score per trial (v4): 94.26 / 94.51 / 96.09 (SD 0.99, range 94.26–96.09)
- GGUF sha256:
c7e0e920a0948849beb7074d8368cbd5ac6825fcd1f2450cf851276325d8a6a5 - Sponsor disclosure: none
Chat template vs upstream
Matches: the chat template in the GGUF we tested is identical to the provider's upstream template after whitespace normalisation.
- GGUF template sha256 (normalised):
4880dc50493b95f47b108a77ff06e25994d2ce8d5e28359d02975d6eefd4aa79 - Upstream reference sha256 (normalised):
4880dc50493b95f47b108a77ff06e25994d2ce8d5e28359d02975d6eefd4aa79· source - Ollama renders this model with its built-in Go renderer
qwen3.5, not the GGUF Jinja template, so the prompt it actually sends is not covered by this hash.
Labels are only “matches / differs / no reference”, never safe/unsafe. Ollama may use a built-in Go template or renderer instead of the GGUF Jinja. Why templates matter: Chat-template backdoors. Not part of the score.