← Model list
Muse Spark 1.1
meta·meta/muse-spark-1.1·GA·Closed·muse series·NEW
Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.
At a glance
Official price$1.25 / $4.25 per 1M
Cache read $0.15 · Meta
Lowest paid$1.25 / $4.25
NanoGPT Gateway · 1× spread
Context1,000,000
⚠ Providers report 1,000,000–1,048,576; the table below is authoritative
Output limit32,000
Capabilities
✓ Reasoning✓ Tool use✓ Structured output✓ Temperature✓ Attachments
⚠ Providers report capability flags inconsistently
Modalities
TextImagePDFVideoAudio
Knowledge cutoff—
Released / updated2026-04-08 / 2026-07-09
Quality & performanceArtificial Analysis · Intelligence Index v4.1 · rep. tier xhigh
Intelligence53.2
Value27
Coding71.3
Agentic39.7
Output speed200 tok/s
TTFT1.48 s
Cost per task$0.2923
Value formulaIQ 53.2 ÷ min blended $2.000 = 27
Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.
Available at 10 providers10 with public prices
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| NanoGPT | Gateway | $1.25 | $4.25 | $0.15 | — | 1,000,000 | 65,536 | |
| MetaOfficial | First-party | $1.25 | $4.25 | $0.15 | — | 1,000,000 | 32,000 | |
| OpenRouter | Gateway | $1.25 | $4.25 | $0.15 | — | 1,048,576 ⚠ | 1,048,576 | |
| Merge Gateway | Gateway | $1.25 | $4.25 | $0.15 | — | 1,048,576 ⚠ | 262,144 | |
| EmpirioLabs AI muse-spark-1-1 | Gateway | $1.25 | $4.25 | $1.00 | — | 1,048,576 ⚠ | 131,072 | |
| Vercel AI Gateway | Cloud | $1.25 | $4.25 | $0.15 | — | 1,048,576 ⚠ | 1,048,576 | |
| DevPass (LLM Gateway) | Gateway | $1.25 | $4.25 | $0.15 | — | 1,048,576 ⚠ | 32,000 | |
| Kilo Gateway | Gateway | $1.25 | $4.25 | $0.15 | — | 1,048,576 ⚠ | 1,048,576 | |
| Pioneer | Gateway | $1.25 | $4.25 | $0.15 | $1.25 | 1,048,576 ⚠ | 131,072 | |
| Abacus | Gateway | $1.25 | $4.25 | $0.15 | — | 1,000,000 | 32,000 |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
Your usage cost
1NanoGPT$330.50
2Meta · Official$330.50
3OpenRouter$330.50
4Merge Gateway$330.50
5Vercel AI Gateway$330.50
6DevPass (LLM Gateway)$330.50
Switch to NanoGPT to save $0.00/mo (0%). The gap is small, so staying on the official channel is fine.
Note: this is a gateway; verify availability and rate limits yourself.
Benchmark11 items
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| SWE-Bench Pro | — | 61.5 | resolve rate | Source ↗ |
| Terminal-Bench | v2.1 | 80 | success rate | Source ↗ |
| DeepSWE | v1.1 | 53.3 | resolve rate | Source ↗ |
| MCP Atlas | — | 88.1 | success rate | Source ↗ |
| JobBench | — | 54.7 | success rate | Source ↗ |
| Toolathlon-Verified | — | 75.6 | success rate | Source ↗ |
| Humanity's Last Exam | variant: with tools | 62.1 | accuracy | Source ↗ |
| OSWorld-Verified | — | 80.8 | success rate | Source ↗ |
| Finance Agent | vv2 | 57.2 | accuracy | Source ↗ |
| CharXiv Reasoning | — | 88.4 | accuracy | Source ↗ |
| BabyVision | — | 76.3 | accuracy | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.
Reasoning control
effort = minimaleffort = loweffort = mediumeffort = higheffort = xhigh
2 / 10 providers expose no reasoning control (reasoning_options: []).
Related models
Muse Glimmer 30Bsame series$0.35 / $1.50Muse Glimmer 30B TEEsame series$0.35 / $1.50Muse Glimmer 30B (TR)same series$2.00 / $5.00DeepSeek V4 Procheaper alternative$0.435 / $0.87DeepSeek V4 Flashcheaper alternative$0.14 / $0.28DeepSeek V4 Flash 0731cheaper alternative$0.479 / $1.44MiniMax-M3cheaper alternative$0.30 / $1.20
Price historyone sample accumulated per data sync
Input listOutput listMin blended