Patchwire

Meta · Released Jul 16, 2026

Muse Spark 1.1

Muse Spark 1.1 is a multimodal reasoning model from Meta, built for agentic tasks. It accepts text, images, video, and PDF documents and returns text, with a 1M-token context window....

ReasoningVision
  • Top 13%SciCode
  • Top 28%Vending-Bench 2
  • Top 28%SimpleQA Verified
  • Top 28%LMCA
Compare in showdown → OpenRouter ↗ All Meta models
Released
Jul 16 2026
Context window
1M 944K out
Input / 1M tokens
$1.25
Output / 1M tokens
$4.25

Where to run it

Providers

Only Meta serves Muse Spark 1.1 on OpenRouter right now.

ProviderInputOutputCache readContextMax outputUptimevs. best
Meta $1.25 $4.25 $0.151M944K100%24h Cheapest

Value

Price vs. performance

7 models score higher for the same price or lesson DeepSWE, e.g. Gemini 3.8 Flash.
11.7%32.5%53.3%74.1% $0.01$0.10$1.00$10.00 GPT-6 Astra (OpenAI) — 74.1% · $10.00/1M in Qwen3.8 Max (0902) (Qwen) — 57.5% · $2.00/1M in Gemini 3.8 Flash (Google) — 73.8% · $0.82/1M in GLM 5.3 Flash (Z.ai) — 63.4% · $0.047/1M in GLM 5.3 (Z.ai) — 69.0% · $0.079/1M in Gemini 3.7 Flash (Google) — 65.5% · $0.75/1M in Grok 4.6 (xAI) — 67.5% · $2.00/1M in Claude Opus 5 (Anthropic) — 73.7% · $5.00/1M in Gemini 3.6 Flash (Google) — 46.7% · $0.75/1M in Kimi K3 (Moonshot AI) — 68.5% · $0.24/1M in GPT-5.6 Luna (OpenAI) — 67.2% · $0.20/1M in GPT-5.6 Terra (OpenAI) — 69.6% · $2.00/1M in GPT-5.6 Sol (OpenAI) — 72.7% · $4.00/1M in Grok 4.5 (xAI) — 53.8% · $2.00/1M in Claude Sonnet 5 (Anthropic) — 53.8% · $2.00/1M in GLM 5.2 (Z.ai) — 43.8% · $0.62/1M in Kimi K2.7 Code (Moonshot AI) — 30.5% · $0.67/1M in Claude Fable 5 (Anthropic) — 69.9% · $10.00/1M in Claude Opus 4.8 (Anthropic) — 59.0% · $5.00/1M in Gemini 3.5 Flash (Google) — 37.4% · $1.50/1M in GPT-5.5 (OpenAI) — 67.0% · $5.00/1M in GPT-5.4 (OpenAI) — 51.8% · $2.50/1M in Gemini 3.1 Pro Preview (Google) — 11.7% · $2.00/1M in Claude Sonnet 4.6 (Anthropic) — 29.9% · $3.00/1M in Muse Spark 1.2 (Meta) — 54.9% · $1.25/1M in Muse Spark 1.1 (Meta) — 53.3% · $1.25/1M in Muse Spark 1.1

Muse Spark 1.1 Other Meta Other labs Value frontier · DeepSWE vs. input price per 1M tokens (log scale) · Drag to pan, pinch or Ctrl + scroll to zoom, double-click to zoom in.

Benchmarks

12 benchmark scores

BenchmarkScoreRankLeader
DeepSWESoftware engineering · Original long-horizon tasks across active open-source repos 53.3% #19of 26 GPT-6 Astra74.1%
GBAEvalSoftware engineering · Build a Game Boy Advance emulator from scratch 7.9% #13of 23 Claude Opus 579.6%
SciCodeSoftware engineering · Research-coding problems from the natural sciences 58.8% #12of 93 Claude Opus 5.566.9%
APEX-AgentsAgents & computer use · Long professional tasks in banking, consulting and law 31.8% #45of 51 Gemini 4 Argon82.4%
Vending-Bench 2Agents & computer use · Run a simulated vending business for a year (final balance) $6,520 #17of 61 GPT-6 Astra$15,515
GDP.pdfWriting & content · Answer questions about real professional PDFs 15.0% #30of 38 GPT-6.1 Sol35.0%
SimpleQA VerifiedWriting & content · Factual accuracy on 1,000 short questions 57.8% #18of 66 Gemini 4 Argon76.0%
WebDev ArenaDesign & visual · People vote on which model builds the better website (Elo) 1542 #35of 88 Claude Opus 5.51820
CritPtReasoning & knowledge · Open-ended, research-level physics problems 15.1% #39of 96 GPT-5.6 Sol32.3%
Chess PuzzlesReasoning & knowledge · Find the single best move in 100 new puzzles 34.0% #44of 70 GPT-6 Astra76.0%
LMCAReasoning & knowledge · Judge the quality of conceptual arguments like experts 49.9% #26of 96 Claude Sonnet 5.568.3%
ProofBenchMaths · Graduate maths proofs verified in Lean 4 39.0% #40of 70 Claude Sonnet 5.5100.0%

Specs

Details

Model ID
meta/muse-spark-1.1
Developer
Meta
Release date
Jul 16, 2026 ·
Context window
1,048,576 tokens
Max output
943,718 tokens
API price
$1.25 / $4.25 per 1M input / output tokens
Providers
1
Input types
text, image, video, file
Reasoning
Yes
Tool calling
Yes
Open weights
No
Free tier
No

Release date, prices and providers from OpenRouter's catalog; buying directly from a provider can cost a little differently. Uptime is the share of OpenRouter requests each provider served over the last day. Benchmark scores are the best result per benchmark from Epoch AI, licensed CC BY 4.0. Rank counts every model tracked here with a score.