Muse Spark 1.2
Meta · Released Aug 2026
Meta's coding-focused model update, co-trained with the Muse Code harness — 82.9% on Terminal-Bench 2.1 and 59.3% on DeepSWE 1.1, second only to Claude Opus 5 on the former and trailing both Opus 5 and GPT-5.6 Terra on the latter. Standard pricing unchanged from 1.1 at $1.25 input / $4.25 output per 1M tokens; contributor tier drops to $0.10/$0.20 in exchange for training-data rights. All published benchmarks remain vendor-run with no independent reproduction. Best for cost-sensitive coding teams willing to test a model explicitly optimized for its own harness.
Is it right for you?
Good for
- Cost-sensitive coding agent workloads — $1.25/$4.25 standard pricing undercuts Claude, GPT, and Gemini flagships by 3-5x
- Muse Code agent integration — co-trained with the harness for best-in-tool performance
- Long-horizon autonomous coding — 1,000+ tool calls over 24 hours, event-log crash recovery, sustained optimization past initial exploration
- Multimodal repo understanding — accepts text, image, video, and PDF inputs for coding context
Not good for
- Production coding accuracy without independent validation — all published benchmarks are vendor-run in Meta's own framework; no third-party reproduction exists yet
- Non-coding or general agentic tasks — coding-focused checkpoint; reasoning, knowledge, math, and multilingual benchmarks remain unpublished
- Open-weight or self-hosting requirements — closed weights, API-only; no Hugging Face weights, no fine-tuning
- High-concurrency production on contributor tier — 60 RPM cap vs 3,000 standard; training-data grant on all contributor traffic
How it performs by task
Code generation
Second on Terminal-Bench 2.1 (3.8 pts behind Opus 5) and Meta's internal coding bench (8.8 pts behind); third on DeepSWE 1.1 behind Opus 5 and GPT-5.6 Terra
Agentic coding
Persistent background agents + worktree isolation; 24-hour kernel optimization demo with 1,000+ tool calls; co-trained with Muse Code harness
Long-horizon tasks
Sustained improvement over 24 hours on GPU kernel optimization; event-log replay on crash prevents lost work and re-prompting
General reasoning
Coding-focused checkpoint; reasoning, knowledge, math, and multilingual benchmark categories remain unpublished — cannot assess non-coding strength
Pricing
Input
$1.25 / 1M tokens
Output
$4.25 / 1M tokens
Context
1M tokens
Benchmarks
No verdict changes yet
The clock starts day one — changes land here as our verdict evolves.
Sources
- Meta — Muse Spark 1.2 evaluation methodology (official PDF)Aug 2026
- Meta AI Research Blog — Muse Code and Muse Spark 1.2 announcement (official)Aug 2026
- Meta Model API — models documentation (official)Aug 2026
- BenchLM.ai — Muse Spark 1.2 profileAug 2026
- VentureBeat: Meta enters AI coding warsAug 2026
- orcarouter.ai — Muse Spark 1.2 explainedAug 2026
- Meta Model API — pricing and rate limits (official)Aug 2026
- Vorp Labs — Muse Spark 1.2 release reviewAug 2026
- explainx.ai — Muse Code BetaAug 2026
Verification log
No verification checks yet
We haven't logged a verification check for this entry. Once a check runs, its history shows here.