🔍 Read the full analysis: Muse Wonder Full Joins AI’s Frontier As The Muse Wonder Charge Returns on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Meta has released Muse Wonder and Muse Wonder Full, with Artificial Analysis reporting Intelligence Index scores of 52 and 56, respectively. The stronger model is available only through Meta’s Muse Code, while the API model costs $4.52 per benchmark task in the analysis—more than several rivals with similar scores.
Meta has released Muse Wonder and Muse Wonder Full, two proprietary models that accept text, images, audio and video, with a one-million-token context window. An Artificial Analysis evaluation published October 12 gave Wonder Full a score of 56 and Wonder a score of 52 on its Intelligence Index; Wonder Full is limited to Muse Code, while Wonder is the version available through Meta’s API.
Artificial Analysis said Wonder Full scored eight points above Muse Spark 1.3, which scored 48, and improved on all ten evaluations in the index. Among 27 configurations compared in the source report, Claude Opus 5.5 at its maximum setting scored higher, at 58, while Claude Sonnet 5.5 at maximum tied Wonder Full at 56. These are benchmark results, not a guarantee of performance on every task.
The two models differ in access and pricing. Muse Wonder is available through Meta’s API and in Muse Code, at listed rates of $4 per million input tokens, $0.20 per million cached input tokens and $20 per million output tokens. Muse Wonder Full is available only to Muse Code users; Meta has not provided a per-token price for it, so an API-style cost comparison is unavailable.
Artificial Analysis reported Wonder Full at 57% on Terminal-Bench 4.0, up from 33% for Spark 1.3, and 71% on AutomationBench-AA. Its GDPval-AA v2.1 result was 1,857 Elo, close to the reported 1,866 for Opus 5.5 and 1,838 for Sonnet 5.5; the source says these results fall within the models’ confidence intervals. Wonder scored 52% on Terminal-Bench and 68% on AutomationBench-AA. The evaluation also recorded hallucination rates of 23% for Wonder Full and 27% for Wonder, compared with 33% for Spark 1.3.
Infographic 01 · Where Meta’s new pair lands
Muse Wonder Full ties for second. Muse Wonder ties GPT-6.1 Sol.
Artificial Analysis scores 27 frontier and open-weight configurations on ten evaluations. Meta’s two new models land near the top of the list. Six weeks earlier, Muse Spark 1.3 scored 48.
Intelligence Index score, 27 configurations
Higher is better. Reasoning effort in brackets, as Artificial Analysis tested it.
Availability: Muse Wonder runs on Meta’s API and in Muse Code. Muse Wonder Full is only available to Muse Code users and has no per-token price.
Source: Artificial Analysis, Muse Wonder analysis (12 Oct 2026) and model leaderboard, read 13 Oct 2026. Intelligence Index v4.3, ten evaluations.
Powered by Thorsten Meyer AI
Strong Scores, Uneven Access and Cost
The release puts Meta’s latest models into closer competition with leading systems on the reported benchmarks, particularly for coding and agent-style work. In Artificial Analysis’s Coding Agents Index, Muse Code with Wonder Full scored 68, tying the listed Claude Code and Sonnet 5.5 configuration. The index measures the model and its software harness together, however, and does not establish how Wonder Full would perform in another coding environment.
For users choosing an API, the cost comparison may matter as much as the score. Artificial Analysis estimated $4.52 for one Intelligence Index task using Muse Wonder, which scored 52. The same report put GPT-6.1 Sol at $0.72 for a score of 52, Gemini 4 Argon at $1.99 for 53, and GPT-6 Astra at $3.26 for 53. Those figures make Wonder a more expensive option in this particular benchmark setup; actual costs will depend on usage and may differ across workloads.
The evaluation’s token analysis points to one contributor: Wonder used 7.69 million input tokens per index task, compared with 6.03 million for Spark 1.3. Artificial Analysis attributed $3.43, or 76%, of Wonder’s estimated $4.52 task bill to input tokens. Its output use was about 54,000 tokens per task. This suggests that input volume, not just the published per-token rate, is important when interpreting the benchmark cost.
As an affiliate, we earn on qualifying purchases.
Six Weeks After Spark 1.3
The release follows Muse Spark 1.3 by about six weeks, according to the source report. That earlier model scored 48 on the Intelligence Index. The latest evaluation records gains for both successors, though Wonder Full and Wonder are not interchangeable products: one is available only within Muse Code, and the other can be purchased through the API.
The reported price comparison also reflects list rates and a specific benchmark method. Meta’s stated API prices for Wonder are $4 per million input tokens and $20 per million output tokens. The report says Spark 1.3 was listed at $1.25 per million input tokens and $4.25 per million output tokens. It compares Wonder’s rates with Claude Opus 5.5’s, while listing $2 per million input tokens and $10 per million output tokens for Sonnet 5.5, Gemini 4 Argon and GPT-6.1 Sol. Task-cost estimates are not the same as a universal price for a user’s workload.
Artificial Analysis also reported Wonder Full at 30 on AA-Omniscience and Wonder at 25, where the assessment rewards correct answers and penalizes confident errors. The report listed hallucination rates of 15% for Gemini 4 Argon, 23% for Wonder Full and 27% for Wonder among the cited higher-scoring models. Those results provide a reason to retain checks in applications where incorrect outputs carry consequences.
“The bill is in the input, not the output.”
— Thorsten Meyer, ThorstenMeyerAI.com
Pricing and Real-World Performance
The report does not give a per-token price for Muse Wonder Full, so its cost per task cannot be compared directly with API models. It is also unclear whether Meta will make the model available outside Muse Code or publish additional pricing information.
Benchmark scores reflect defined tests and do not establish how either model will perform across every customer workload. The source material does not provide details sufficient to independently assess all test conditions, or explain how the reported token use and task costs might vary with different prompts and applications. Real-world costs and accuracy remain workload-dependent.
Watch for Wider Access and Usage Data
The next useful developments will be any Meta announcements on Wonder Full access or pricing, and further evidence on how both models perform in practical deployments. API users can compare Wonder’s published rates with their own token volumes and task requirements rather than treating the benchmark’s $4.52 estimate as a fixed charge.
Artificial Analysis’s published evaluation provides the current comparison in the source material. No additional release date, pricing change or expansion of availability is confirmed there. Users considering the models will need to weigh the reported benchmark gains against cost, access limits and the need to verify outputs.
Key Questions
What did Meta release?
Meta released Muse Wonder and Muse Wonder Full, proprietary models with a one-million-token context window and support for text, images, audio and video inputs.
How did the models score?
Artificial Analysis reported Intelligence Index scores of 56 for Wonder Full and 52 for Wonder. The figures are benchmark results and do not predict performance on every task.
Can customers access Muse Wonder Full through the API?
Not according to the source report. Wonder Full is available only to Muse Code users; Muse Wonder is the model available through Meta’s API.
How much does Muse Wonder cost?
Meta’s listed API rates are $4 per million input tokens, $0.20 per million cached input tokens and $20 per million output tokens. Artificial Analysis estimated $4.52 for one task in its Intelligence Index evaluation, a benchmark-specific figure rather than a universal task price.
What remains unknown about the release?
The source does not give Wonder Full a per-token price or indicate whether it will become available outside Muse Code. Performance and costs on real customer workloads also remain uncertain and dependent on use.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
