Muse Spark 1.2
MetaProprietaryReasoningVerifiedReleased Aug 5, 2026Input: text, image, video, file
Overall score
65.9
Rank 34 of 222 · 59.8–72.0
Context window
1M
1,000,000 tokens
Price per 1M tokens
$1.25 in · $4.25 out
Source: BenchLM
Blended price
$2.00
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Reasoning79.3
- 2Agentic & tool use#2675.6
- 3Knowledge#3572.2
- 4Coding#4062.9
Benchmark results (6)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Coding
| DeepSWE | 59.3 | #37 of 50 |
| FrontierSWE v2 | 12.0 | #19 of 20 |
| VulcanBench v3 | 87.0 | #4 of 14 |
Agentic & tool use
| Terminal-Bench 2.1 | 82.9 | #23 of 68 |
| MazeBench | 0 | #20 of 42 |
Multimodal
| Design Arena Website | 1313 | #10 of 91 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| List price (BenchLM) | $1.25 | $4.25 | $0.15 |
| OpenRouter meta/muse-spark-1.2 | $1.25 | $4.25 | $0.15 |
Meta Model API lists muse-spark-1.2 at $1.25 input, $0.15 cached input, and $4.25 output per million tokens. Meta separately lists muse-spark-1.2-contributor, whose data may be used to improve Meta products, at $0.10 input, $0.002 cached input, and $0.20 output per million tokens; BenchLM stores the standard non-contributor SKU as the headline rate.
Arena ratings
Human-preference ratings from LMArena. Best-rated variant shown.
| Text(xHigh) | 1484 | #15 · 8.2K votes |
| WebDev (coding)(xHigh) | 1534 | #51 · 2.2K votes |
| Vision(xHigh) | 1304 | #17 · 1.9K votes |
Speed
Measured by us through OpenRouter (median of trials).
Not measured yet.
Related models
Same family
Similar score
- Muse Spark 1.166.1 · $2.00
- DeepSeek V4.1 Flash 101765.7 · $0.26
- Claude Opus 4.765.5 · $10.00
- GPT-6 Luna66.6 · $0.20
- GPT-5.2 Pro66.7 · $57.75