Clairic-flash
Licence: Apache 2.0
The compact member of Cloudflare's extraction family, released with full open weights Nov 18, 2026 under Apache 2.0: a 9B multimodal System One model that turns a text or JSON state plus optional images into typed, scored fields — extraction, estimates and forecasts. Built on Qwen3.5-9B (per its Hugging Face card, read first-hand by this desk). Memory by arithmetic: about 18 GB at BF16 — one consumer 24GB card. Hosted comparator: Workers AI at $0.10 per million input tokens, no output charge. Specialised extraction model, not a general chat model; all benchmarks are Cloudflare's internal evaluations.
What it needs
How we work this out: memory = the model's total parameters × bytes per weight — 2 bytes at full precision (FP16), 1 byte at 8-bit, half a byte at 4-bit. Weights only — add room for the conversation memory (KV cache) and the serving software. A GPU counts as running a build when its memory covers the weights plus about 20% headroom. Real figures move with context length and software.
GPUs that can run it
Ranked by how well the model runs, then by price.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 24GB, with room left for the conversation memory.
24GB fills up fast on big models. No NVLink, so two cards can't pool memory — and it is a gaming card at heart.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 48GB, with room left for the conversation memory.
An older generation — no FP8, so the newer speed tricks for serving don't apply.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 32GB, with room left for the conversation memory.
No NVLink, so two cards can't pool memory. A consumer card — driver and ECC behaviour matter for production serving.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 95GB, with room left for the conversation memory.
This is the Max-Q version: it runs at lower power and is slower than the full card.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 48GB, with room left for the conversation memory.
70B at FP16 does not fit, and no flag fixes that. Weak at double-precision maths; no NVLink.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 48GB, with room left for the conversation memory.
Weak at double-precision maths — not the card for scientific computing. No NVLink.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 80GB, with room left for the conversation memory.
An older generation — no FP8, so the newer speed tricks for serving don't apply.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 96GB, with room left for the conversation memory.
No NVLink scale-out; strong all-round, but the Max-Q version (listed separately) runs slower.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 40GB, with room left for the conversation memory.
The 40GB fills up quickly on larger models.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 80GB, with room left for the conversation memory.
No NVLink, and lower power than the SXM — slower when cards must work together.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 80GB, with room left for the conversation memory.
Pays off only when the card stays busy; on bursty work you pay H100 rent for idle silicon.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 94GB, with room left for the conversation memory.
Sold as linked pairs, not as a lone card.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 80GB, with room left for the conversation memory.
One provider's label, published alongside its plainer “H100 PCIe” — check which card you are actually booking.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 141GB, with room left for the conversation memory.
More card than most jobs need; the main trap is paying for it on small work.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 80GB, with room left for the conversation memory.
Pays off only when kept busy — avoid it for bursty development work.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 192GB, with room left for the conversation memory.
Frontier training money. If your job fits on anything cheaper, a B200 is budget burned for a spec sheet.
Runs Clairic-flash at full precision — its 18GB of weights fit this card's 96GB, with room left for the conversation memory.
Its processor is ARM, not the usual x86 — some software needs changes to run on it.
The cheapest verified way to run it
RTX 4090 at RunPod Community — $0.34/hr — runs it well, at full precision.
Launch RunPod Community ↗ See the RTX 4090 →
Checked Nov 21, 2026
Select a GPU See these GPUs on the prices board, ranked for this model.