Hi!
DIT Model Radar tracks new AI models daily — real models or confirmed codenames only, no tools, no hype.
Meta's Muse Wonder lands on the Meta Model API at $4 / $20 per million tokens, the same price as Claude Opus 5.5. By Meta's own numbers it beats Opus 5.5 on long-horizon coding, 53.1 to 46.3 on SWE-Marathon, but it scores 52 to Opus 5.5's 58 on the independent Artificial Analysis index, and the stronger Wonder Full is not sold on the API.

What changed: Muse Wonder (model ID muse-wonder-1.0) and Muse Wonder Full open Meta's new Wonder series, trained at Meta's Prometheus data center, with text, image, audio and video input, a 1M-token context window and no downloadable weights. At max effort Muse Wonder scores 78.8 on DeepSWE 1.1 and 53.1 on SWE-Marathon against Opus 5.5's 74.2 and 46.3, but trails it on Terminal-Bench 4.0, 52.0 to 59.6. Muse Wonder Full posts 79.0 and 57.5 and ties Claude Sonnet 5.5 at 56 on the Artificial Analysis Intelligence Index, still behind Opus 5.5 at 58.
- Muse Wonder: $4 / $20 per MTok input/output, the same as Claude Opus 5.5
- Cached input: $0.20 per MTok
- Muse Wonder Full: no API price, included in Muse Code Power Usage ($50/month) and Maximum Usage ($200/month) plans
Meta just matched Opus 5.5 on price, not on the scoreboard — and the version that comes closest isn't for sale on the API.
![]() |

Top AI models. Lower costs.
Access GPT, Claude, Gemini, and more through one OpenAI-compatible API—and save up to 70% on AI API costs. Create your DIT account, get your API key, and start building.

