Compare AI models

Specifications
Parameters
—1.05T
Context window
1M1M
API pricing
Input price
$1.25$0.70
Output price
$4.25$2.10
Cached input price
$0.15$0.07
Cheapest input
$1.25Meta$0.70Mistral
Cheapest output
$4.25Meta$2.10Mistral
$463Meta list price$245Mistral list price
Benchmarks
BullshitBench v2
50%80%
DeepSWE 1.1
59.3%61.7%
SWEAtlas CodeBase QnA
46.2%59.4%
AutomationBench
38.2%59.9%
Overview
CompanyMetaMistral
Release dateAug 5 2026Oct 6 2026
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Mistral Large 4 leads Muse Spark 1.2 on 4 of the 4 benchmarks they both report (BullshitBench v2, DeepSWE 1.1, SWEAtlas CodeBase QnA, AutomationBench). Mistral Large 4 is cheaper on both input and output: $0.70 vs $1.25 per million input tokens, and $2.10 vs $4.25 per million output tokens. Muse Spark 1.2 shipped 62 days before Mistral Large 4, so benchmark comparisons should account for the intervening progress.

Context windows are 1M (Muse Spark 1.2) vs 1M (Mistral Large 4). Muse Spark 1.2 is closed, while Mistral Large 4 is open weight.

On BullshitBench v2, Mistral Large 4 leads at 80% vs Muse Spark 1.2 at 50%. On DeepSWE 1.1, Mistral Large 4 leads at 61.7% vs Muse Spark 1.2 at 59.3%. On SWEAtlas CodeBase QnA, Mistral Large 4 leads at 59.4% vs Muse Spark 1.2 at 46.2%. On AutomationBench, Mistral Large 4 leads at 59.9% vs Muse Spark 1.2 at 38.2%.