models · allenai

Olmo 3 32B Think

Olmo 3 32B Think — catalogued from third-party evaluation data; no weights held by Sovereign Frontier.

Compare Verify

What this registry has checked

Quarantine. We hold the publisher’s own published hash for these weights and have recorded it against this entry. We have not downloaded and hashed the weights ourselves, so the hash is the publisher’s claim, not our measurement.
Methodhf-published-sha256

What is recorded

Publisherallenai
OriginUnited States
LicenceApache-2.0
Parameters32,233,522,176 (32.2B)
Context window65,536 tokens
Modalitytext->text
Architectureolmo3
Layers64
Hidden size5120
AttentionGQA (40 heads, 8 KV)
Native precisionBF16
Pulls recorded here6
Downloads reported upstream9,502

Reported scores

MMLU-Pro75.9
GPQA Diamond61
AIME 202573.7
LiveCodeBench67.2
IFBench49.1
Humanity's Last Exam6.4

Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.

Measured on real hardware

None of these runs are ours. Each figure is the reporter’s, linked to their own evidence, under their own trust label. The number shown is single-stream decode — one user, concurrency 1 — because aggregate throughput across many users is a different quantity and reads as far faster than anyone will see.

Reported to local-ai-registry

HardwareEngineBuild measuredDecode, 1 streamLabelReported by
GeForce RTX 3090llama.cppQ4_K_M23.1 tok/s @ 512candidatelocalmaxxing
GeForce RTX 3060 ×2 devicesllama.cppQ4_K_M16.4 tok/s @ 512candidatelocalmaxxing

From local-ai-registry (MIT), commit 124959d: 2 runs across 2 machines, 0 validated — model revision and runtime pinned, launch accepted — and 2 candidate, their label for useful evidence without a reproducible-launch promise.

Elsewhere in the registry

Registry record as of 2026-08-13. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.