Muse Spark 1.3 vs Muse Spark 1.1
Muse Spark 1.3 wins on Overall, Coding, Multimodal. Every numeric field is compared below, with a worked monthly-cost example and a pick rule for each use case.
Verdict at a glance
Choose Muse Spark 1.3 if you…
- #1 DeepSWE long-horizon coding
- Open weights, self-hostable
- Extremely low cost
- Fast
- Best for: Coding, Open-source deployment, Budget
Choose Muse Spark 1.1 if you…
- Completely free open weights
- By Meta
- Self-hostable
- Multimodal
- Best for: Open research, Local deployment, Experimental
Head-to-head aggregate scores
Scores are 0–100, aggregated from public benchmark information and independently weighted across three leaderboards. Rank is out of 26 tracked models.
Aggregated from public sources and independently weighted; methodology on the Terms page. Scores within 3 points are treated as statistically tied.
Specs & pricing — every field side by side
List API prices in USD per 1M tokens. The highlighted cell is the stronger value/capability on that row.
| Dimension | Muse Spark 1.3 | Muse Spark 1.1 | Verdict |
|---|---|---|---|
| Vendor | Meta (US) | Meta (US) | Same vendor |
| Released | 2026.09 | 2026.08 | Muse Spark 1.3 is newer |
| Overall (rank) | 64 · #22 | 61 · #26 | Muse Spark 1.3 +3 |
| Coding | 88 · #5 | 59 · #25 | Muse Spark 1.3 +29 |
| Multimodal | 76 · #17 | 75 · #19 | Muse Spark 1.3 +1 |
| Context window | 256K | 256K | Tie |
| Max output | 64K | 64K | Tie |
| Effective-context | 85 | 80 | Muse Spark 1.3 more reliable |
| Input $/1M | $1.25 | Free | Muse Spark 1.1 cheaper |
| Output $/1M | $4.25 | Free | Muse Spark 1.1 cheaper |
| Cache discount | none | none | Tie |
| Speed | ~70 tok/s | Hardware-dependent | Muse Spark 1.3 faster |
| TTFT | 0.4s | Hardware-dependent | Muse Spark 1.3 snappier |
| Function calling | 72 | 65 | Muse Spark 1.3 ahead |
| Refusal rate | ~4% | ~4% | Tie |
| English | 84 | 78 | Muse Spark 1.3 |
| Chinese | 62 | 60 | Muse Spark 1.3 |
| Modalities | text, image | text, image | Same |
| Open weights | Yes | Yes | Both open |
| Fine-tuning | Yes | Yes | |
| Free tier | Open weights free to self-host; hosted API at $1.25/$4.25 per 1M | Fully free (model weights) | |
| SOC2 / no-train | no / yes | no / yes | |
| Private deployment | Yes | Yes | Both support it |
Fields drawn from vendor public documentation and the Modelspectra 26-model dataset; speed varies with network, concurrency and prompt length. Verify current pricing before purchase.
Dimension-by-dimension analysis
Reasoning & overall intelligence
Muse Spark 1.3 leads the overall aggregate by 3 points (64 vs 61). That is a meaningful, not marginal, edge on hard, multi-step reasoning, while Muse Spark 1.1 remains a strong generalist that is not out of its depth on routine work.
Agentic coding
This is a decisive gap: Muse Spark 1.3 scores 88 against 59. On multi-file edits, SWE-style tickets and long-horizon agent loops Muse Spark 1.3 needs fewer correction turns; Muse Spark 1.1 is still competent for scripts and assisted completion but trails as task complexity rises.
Multimodal
Muse Spark 1.3 leads multimodal 76 vs 75. Neither emits native video, so the comparison is about parsing images and documents, not generation.
Speed & latency
Muse Spark 1.1 is self-hosted, so its speed depends on your hardware; against a managed API like Muse Spark 1.3 (~70 tok/s, 0.4s TTFT) compare on your own infrastructure before assuming latency.
Context: window vs usable recall
Nominal windows are 256K for Muse Spark 1.3 and 256K for Muse Spark 1.1. Effective-context scores point the same way as window size — Muse Spark 1.3 is ahead on usable recall (85 vs 80), so prefer it for long-document work where details cannot be missed.
Price & total cost
Muse Spark 1.1 is free open weights (you pay only for the infrastructure you run it on), while Muse Spark 1.3 is a paid API at $1.25/$4.25 per 1M tokens. The real comparison is total cost of ownership — GPU/ops against a managed bill — not list price alone.
Chinese vs English
English: Muse Spark 1.3 84 vs Muse Spark 1.1 78. Chinese: 62 vs 60. Both are US-based models; for Chinese-first workloads also compare domestic models on the leaderboard.
Tool use & ecosystem
Function-calling score: Muse Spark 1.3 72 vs Muse Spark 1.1 65, so Muse Spark 1.3 has the edge on structured tool use. Fine-tuning is available from Muse Spark 1.3 and Muse Spark 1.1. Factor in existing SDK/plugin familiarity — switching cost often outweighs a few-point tool-use gap.
Cost worked example — same workload, real token math
Assume a production workload of 100M input + 30M output tokens per month, with 90% of input tokens served from cache. Figures use public list prices.
Illustrative model; your input/output mix and cache-hit ratio change the result. Prices are list rates before any enterprise agreement.