Models › Muse Spark 1.3
Meta • Launched September 2026 for agentic coding
Muse Spark 1.3 is Meta’s agentic model for coding work
Starting provider API rate; context tiers apply
See what Spark 1.3 does best, then review its limits, benchmarks, and pricing below
Muse Spark 1.3 is Meta’s agentic model for coding work
Every Spark 1.3 result below comes from a named source and is labeled by protocol, so each score reflects how it was measured
| Benchmark | Score | Evaluation details | Source |
|---|---|---|---|
GPQA Diamond The highest-quality subset of Graduate-Level Google-Proof Q&A, designed to test expert-level scientific reasoning in biology, physics, and chemistry. |
94% Accuracy |
Version: GPQA Diamond 198 · Split: 198 Diamond questions · Harness: Artificial Analysis independent evaluation · Reasoning effort: xhigh | Methodology Artificial Analysis Independently reproduced |
Humanity's Last Exam A 2,500-question expert-level benchmark spanning dozens of academic fields. Tool-assisted and no-tools results are separate protocols. |
47% Accuracy |
Version: Humanity's Last Exam · Harness: Artificial Analysis independent evaluation · Reasoning effort: xhigh | Methodology Artificial Analysis Independently reproduced |
Terminal-Bench 2.1 A terminal-based agent benchmark of 89 tasks, graded on whether an isolated sandbox reaches the correct final state. |
85% Task success rate |
Version: Terminal-Bench 2.1 · Split: 89 terminal tasks · Harness: Artificial Analysis independent evaluation · Reasoning effort: xhigh | Methodology Artificial Analysis Independently reproduced |
GDPval-AA v2 A knowledge-work benchmark scoring long, realistic professional tasks against expert-produced deliverables across dozens of occupations. |
1,709 Elo score |
Version: GDPval-AA v2 · Harness: Artificial Analysis independent evaluation · Scoring: Pairwise Elo comparison against expert-produced work · Reasoning effort: xhigh | Methodology Artificial Analysis Independently reproduced |
Tau3-Bench Banking A domain-specific agentic benchmark that tests multi-turn customer-service and banking workflows requiring tool use and policy compliance. |
47% Task success rate |
Version: Tau3-Bench Banking · Harness: Artificial Analysis independent evaluation · Reasoning effort: xhigh | Methodology Artificial Analysis Independently reproduced |
Arena rankings capture how people vote between anonymous responses. They are reported separately from task benchmarks and are not used as substitutes for missing benchmark results
No exact Arena match is available for this model. Similar names and provider variants are not merged automatically.
Muse Spark 1.3 runs a 1,000,000-token context window with image input, parallel tool calling, and adjustable reasoning, priced below
| Model | |
| Provider | Meta |
| Release date | Sep 2, 2026 |
| Catalog added | Sep 22, 2026 |
| Context window | 1,000,000 tokens |
| Maximum output | Not available in reviewed sources |
| Knowledge cutoff | Not available in reviewed sources |
| License | Proprietary |
| Capabilities and API | |
| Input types | Image, Text |
| Output types | Text |
| Reasoning | Supported (xhigh default; limited-preview max variant) |
| Provider endpoint tool parameters | Accepts parallel tool calling |
| Base provider API input | $1.25 / 1M tokens |
| Base provider cached input | $0.15 / 1M tokens |
| Base provider API output | $4.25 / 1M tokens |
Provider token rates are context-dependent. The base rates apply below the listed prompt thresholds; the matching tier applies at or above each threshold. These are reference provider API rates, not Pyxa AI checkout charges — access in Pyxa AI follows the allowances of your Pyxa AI plan.
Each Spark 1.3 benchmark and specification links to its original source, so you can check where every figure comes from
Open another model’s detail page or compare it directly with Muse Spark 1.3. Suggestions come from the current catalog
Put Muse Spark 1.3 to work on your coding and agentic tasks inside Pyxa AI