Last checked 31 Oct 2026
GLM 5.3 vs Muse Spark 1.3
GLM 5.3 is billed by Mistral. Muse Spark 1.3 is billed by Meta. A pair is two models from different providers whose higher published input rate is at most 1.5 times the lower one, and whose higher published output rate is at most 1.5 times the lower one. Closer rates are listed first. Each model is in at most 2 pairs, and the list stops at 12. The figures below are the rates stored on those two rows, the differences between them, and the cost of one sample workload. This page does not score quality.
Stored rates
GLM 5.3 (Mistral, Standard) and Muse Spark 1.3 (Meta, Standard). Provider pages: Mistral and Meta.
| Field | GLM 5.3 | Muse Spark 1.3 |
|---|---|---|
| Input / 1M | $1.40 | $1.25 |
| Cached input / 1M | $0.14 | $0.15 |
| Output / 1M | $4.40 | $4.25 |
| Rate | Standard | Standard |
| Context | 1M | 1,048,576 |
| Max output | 128k | n/a |
| Status | Active | Active |
| Modalities | text → text | text, image, video, audio, pdf → text |
| Released | 28 Sep 2026 | n/a |
| Checked | 31 Oct 2026 | 31 Oct 2026 |
| Per request | $0.00632 | $0.0059 |
| Per month | $189.60 | $177.00 |
Sample workload
The sample is the calculator's starting workload: 2,000 input tokens and 800 output tokens per request, 1,000 requests a day. A month in the calculator is 30 days, so that is 30,000 requests. The cost uses the stored input and output rates only. Cache hits, batch discounts, long-context surcharges, peak hours, and prices scheduled after an introductory window are not included. The rate label and the notes below name those other published rates.
GLM 5.3 is $0.00632 per request and $189.60 per month ($84.00 input and $105.60 output over the month). Muse Spark 1.3 is $0.0059 per request and $177.00 per month ($75.00 input and $102.00 output over the month).
Muse Spark 1.3 costs $12.60 less per month than GLM 5.3 on this sample, using the stored rates above.
Open GLM 5.3 in the calculatorOpen Muse Spark 1.3 in the calculator
Notes from the sources
GLM 5.3
Standard
Third-party open-weight model from Z.ai, hosted by Mistral without modifications. The model card lists a 1M context window, 128k max output, text input, text output, and status GA. The changelog dates general availability on Mistral to 28 September 2026. Also listed on the Mistral pricing page.
Muse Spark 1.3
Standard
Standard-mode rate. A Contributor mode (muse-spark-1.3-contributor) is priced at $0.10 input, $0.002 cached input, and $0.20 output per 1M tokens in exchange for Meta using the content for training. That Contributor rate is not the headline price. Meta notes audio input on 1.3 is not fully supported. A release day was not on the pages checked.