Llama 4 Maverick

Meta's current open-weight flagship (mixture-of-experts, large context), replacing the 2024-era Llama 3.3 70B. Newly added to the lineup; playground benchmarking is in progress, so no completion metrics yet. The prior Llama generation favoured completion over efficiency.

Avg tokens per run: 38,679
Avg duration: 13.9s
Turns to checkout: 5

Start a run with Llama 4 Maverick Compare

Shopping score

20/100Weak

Scores come from runs on UCP stores, not estimates: every figure is a tool call, a checkout attempt or a store’s answer.

Shopping score breakdown

Checkout rate40%2.2%
Cart rate20%6.5%
Search rate10%34.8%
Turn efficiency15%88%
Token efficiency15%10%

65.2% of runs failed with errors.

Top stores

Development stores are hidden.

Store	Checkout	Cart
store hidden	0%	33.3%
ucp.travel	0%	0%
singon.com.np	0%	0%
everlane.com	0%	0%
allbirds.com	0%	0%
store hidden	0%	0%

Known issues

No known issues documented for this model yet.

Token usage

Avg per run: 38,679
Average of all models: 77,681
Against the average: −50%
Median (p50): 13,985
p90: 71,034

Prompt and completion

Prompt 38,471 (99%)Completion 208 (1%)

Daily average tokens per run, last 30 days

Oct 14Total 6,324Prompt 6,290Completion 34

Range

Min: 2,093
p25: 9,018
p50: 13,985
p75: 27,203
p90: 71,034
Max: 418,874

Test Llama 4 Maverick on your store

Start a run to see how Llama 4 Maverick handles your store’s checkout, call by call.

Start a run