SMILES: caffeine with three CF3 groups, by deepseek-v4.1-flash
deepseek-v4.1-flash · round 1 · seed 1$0.02SVGmolecule.smi
molecule.smi · 2 lines
O=c1c2c(ncn2C(F)(F)F)n(C(F)(F)F)c(=O)n1C(F)(F)F
gallery/Caffeine with three CF3 groups/DeepSeek 4.1 Flash, run 1

The exact molecule on the first render, and it corrected the prompt pack's formula.

modeldeepseek/deepseek-v4.1-flashtasksmiles-fluorocaffeine@v1date2026-09-22rounds1 of 8 allowedcost$0.0172thinkingmediumtokens56,741 in, 9,084 out, of which 7,729 reasoningtime1 min 44 s
Parsesdecided by a number, from the filepass
C8HF9N4O2decided by a number, from the filepass
Three N-CF3, no N-methyldecided by a number, from the filepass
The right moleculedecided by a number, from the filepass

There is no post about this one, since only some results need one, so the picture, the source and the scoring are here instead.

its other runs
Caffeine with three CF3 groups, by DeepSeek 4.1 Flash, run 2run 2 · 1r · $0.0135The exact molecule on the first render.Caffeine with three CF3 groups, by DeepSeek 4.1 Flash, run 3run 3 · 1r · $0.0219The exact molecule on the first render.
same prompt, other modelsall 3 runs on Caffeine with three CF3 groups →
the prompt it was given
Give the SMILES for caffeine with all three N-methyl groups replaced by trifluoromethyl groups.
Frozen at v1. This model got this and nothing else.
ground truth

Caffeine's skeleton with a CF3 on N1, N3 and N7, which is C8HF9N4O2. The prompt pack this came from gives C8H3F9N4O2, which is wrong: swapping the methyls takes all nine hydrogens with them, and only the C8 hydrogen stays. Decided with RDKit by an exact canonical match, because fluorinated isocaffeine has the same formula with a CF3 on the wrong nitrogen. How it is scored, and this run's files on GitHub.