#72Mistral· released 2025-12-09

Mistral: Devstral 2 2512.

Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding. It is a 123B-parameter dense transformer model supporting a 256K context window. Devstral 2 supports exploring...

View on OpenRouter ↗
Context
262K
Input $/1M
$0.40
Output $/1M
$2.00
Modality
text · file
Arena Elo
1545
LMArena — human blind votes
AA Index
63.0
Artificial Analysis composite
LiveBench
77.1
Contamination-resistant
GPQA
76.4%
Grad-level science reasoning
MMLU-Pro
78.1%
Broad knowledge
SWE-bench
68.8%
Real GitHub bugs (Verified)
Terminal-Bench
58.7%
Agentic shell tasks
ARC-AGI
20.9%
Abstract reasoning
Aider Polyglot
78.2%
Multi-language code editing
Speed
76 tok/s
Median output speed

Profile

Mistral: Devstral 2 2512 is Mistral's entry at position #72 on the current aggregated leaderboard. Its strongest signals are real-world software engineering, unattended agentic work. Every figure below is aggregated from the public sources listed on the homepage and refreshed on each pipeline run.

The score spread is comparatively complete for this model, but the standing advice still applies: treat gaps of a few points as ties and validate on your own workload before committing.

How this model's position is computed is documented in the methodology; what each benchmark actually measures is in the benchmark guide. See a figure that disagrees with its source? Report it — accepted corrections apply on the next refresh.

Nearby on the leaderboard