ministral-3:8b
Family: ministral-3 · 8B parameters · 1 run(s)
Fits in 16 GB VRAM · ollama 0.40.6 · Q4_K_M
Measured VRAM at 32K context (KV cache included, ollama ps on the RTX 3090): 9.8 GB · previous tier: 8 GB (results v0.1.1). Moved to the 16 GB tier in results v0.1.2 (2026-10-26) because it needs more than 8 GB at 32K context; score unchanged. Tier rule
56.3/100 · scoring v4
95% task-bootstrap CI: 50.0–62.7 · ≈ statistically tied with other models in tie group 3 of this tier · rank #9 in tier, rank range under weight sensitivity #8–#10 (5th–95th percentile; min #7, max #10) · v3: 55.4 · mention 6.7% (3/45) · how
Plain-language summary
Mixed — useful for some agent tasks, shaky on others.
Good for
- —
Weak at
- multi-step plans
Category breakdown
| JSON | Schema | Tool | Multi-step | Injection (useful&safe) | tok/s | p50 s/task | p95 s/task |
|---|---|---|---|---|---|---|---|
| 77 | 63 | 65 | 37 | 56 | — | 11.3 | 14.2 |
Prompt injection (15 tasks × 3 trials): useful and safe 55.6% (task completed and not compromised; counts in the score) · utility under attack —% (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): 6.7% (3/45); they do not lower the score. v3 useful and safe: 51.1%. v2.1 used injection resistance 93.3 × tool-call rate . Definitions
- Run:
20261112-ministral-3-8b-6eb673· 2026-11-12T07:24:30+00:00 - GPU: NVIDIA GeForce RTX 3090 (24 GB) · Linux 6.8.0-57-generic
- Host:
· kind not recorded · measured · 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)
- GGUF sha256:
3ea32f882c990baebcfe6ab5cda2e3167d54ab16d7937c0b9aa3db11aba01574 - Sponsor disclosure: none
Chat template vs upstream
No reference
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.