EchoLikeUpdated Oct 23, 9:23 AM Bogotá
AI models on the OpenCode Go plan
Every model the OpenCode Go plan serves (a flat $10 or $40 a month, not per token): how many requests each one allows per 5-hour window and month, and how fast it answered in our own timed runs. Plan limits come from opencode.ai/go; the speed numbers we measured ourselves on 2026-10-22.
live · Refreshed every 24 hours · Updated Oct 23, 9:23 AM Bogotá · 66% of cells are filled (94 of 280 have no value yet).
Hidden by default: 2 more columns, including Failed or empty answers (%), Notes. Use “All columns” or the chooser to see them.
1 entry has no checked fact yet, so it is not shown here until it does.
Columns:
Choose…
Name is always shown: the chooser lists the other columns.
| Model | Plan page | Requests per 5 hours on the OpenCode Go plan | Monthly usage value on the OpenCode Go plan ($) | Requests per 5 hours on the OpenCode Go Plus plan | Monthly usage value on the OpenCode Go Plus plan ($) | Answer speed (tokens per second) | Time to first token (milliseconds) | Failed or empty answers (%) | Notes |
|---|---|---|---|---|---|---|---|---|---|
| Kimi K3 | — | 110 | $15 | 440 | $60 | — | — | — | — |
| Qwen3.8 Max | — | 160 | $15 | 640 | $60 | — | — | — | — |
| Grok 4.7 | — | 169 | $15 | 676 | $60 | — | — | — | — |
| GLM-5.3 | — | 220 | $15 | 1760 | $120 | — | — | — | — |
| Step 5 Preview | — | 430 | $15 | 1720 | $60 | — | — | — | — |
| Gemini 4 Argon Contributor | — | 440 | $60 | 1760 | $240 | — | — | — | — |
| GLM-5.2 | — | 880 | $60 | 2640 | $180 | — | — | — | — |
| DeepSeek V4 Pro | — | 1050 | $15 | 4200 | $60 | — | — | — | — |
| Kimi K2.7 Code | — | 1350 | $60 | 4050 | $180 | — | — | — | — |
| Hy4 preview | — | 1350 | $30 | 5400 | $120 | — | — | — | — |
| GPT 5.6 Luna | — | 2050 | $15 | 8200 | $60 | — | — | — | — |
| MiniMax M3 | — | 3200 | $60 | 9600 | $180 | — | — | — | — |
| MiMo-V2.6-Pro | — | 3250 | $15 | 13000 | $60 | — | — | — | — |
| MiMo-V2.5-Pro | — | 3250 | $15 | 13000 | $60 | — | — | — | — |
| MiniMax M3.1 Flash | — | 3880 | $20 | 15500 | $80 | 90.6 | 1859 | 0% | — |
| Claude Haiku 5.5 | — | 3850 | $15 | 15380 | $60 | — | — | — | — |
| Qwen3.7 Plus | — | 4300 | $60 | 12900 | $180 | — | — | — | — |
| GPT 6 Luna | — | 4230 | $15 | 16920 | $60 | — | — | — | — |
| Hy3 | — | 4300 | $60 | 17200 | $240 | — | — | — | — |
| Qwen3.8 Flash | — | 5400 | $30 | 16200 | $90 | 36.2 | 1660 | 0% | — |
| GLM-5.3-Flash | — | 6320 | $60 | 18960 | $180 | 83.3 | 1407 | 0% | — |
| DeepSeek V4.1 Flash | — | 34670 | $80 | 69330 | $160 | 90.4 | 3213 | 0% | — |
| LongCat-2.0 | — | 11400 | $60 | 45600 | $240 | — | — | — | — |
| MiMo-V2.6-Flash | — | 30100 | $60 | 60200 | $120 | 20.9 | 4614 | 0% | — |
| MiMo-V2.5 | — | 30100 | $60 | 60200 | $120 | — | — | — | — |
| Muse Spark 1.3 Contributor | — | 45300 | $60 | 90600 | $120 | 165.5 | 3172 | 0% | — |
| Muse Spark 1.2 Contributor | — | 45300 | $60 | 90600 | $120 | — | — | — | — |
| Beam Free | — | — | — | — | — | 184.1 | 1152 | 0% | Free with unlimited requests for a limited time on the OpenCode Go plan. |
How this data is built · What each column means · Column names and values are in English: the directory keeps one data standard, so a value means the same thing in every language version.
Where these values come from
Per column: how many values, from which sites, when they were last checked, and one full link you can open. Every value in the table links to its own source.
- Requests per 5 hours on the OpenCode Go plan: 27 values from opencode.ai (newest check Oct 22, 4:52 PM Bogotá) · 15 checked by a person, 12 found automatically. https://opencode.ai/go
- Monthly usage value on the OpenCode Go plan ($): 27 values from opencode.ai (newest check Oct 22, 4:52 PM Bogotá) · 15 checked by a person, 12 found automatically. https://opencode.ai/go
- Requests per 5 hours on the OpenCode Go Plus plan: 27 values from opencode.ai (newest check Oct 22, 4:52 PM Bogotá) · 15 checked by a person, 12 found automatically. https://opencode.ai/go
- Monthly usage value on the OpenCode Go Plus plan ($): 27 values from opencode.ai (newest check Oct 22, 4:52 PM Bogotá) · 15 checked by a person, 12 found automatically. https://opencode.ai/go
- Answer speed (tokens per second): 7 measured by us on this machine.
- Time to first token (milliseconds): 7 measured by us on this machine.
- Failed or empty answers (%): 7 measured by us on this machine.
- Notes: 1 values from opencode.ai (newest check Oct 22, 4:52 PM Bogotá) · 1 found automatically, none checked by a person yet. https://opencode.ai/go
What each column means
- Model: The model's name, as the OpenCode Go plan lists it. Source: entered by hand.
- Plan page: The page where the plan states this model's limits (opencode.ai/go). Source: entered by hand.
- Requests per 5 hours on the OpenCode Go plan: How many requests this model allows in a 5-hour window on the OpenCode Go plan ($10 a month), as the plan page states it. Source: the project's own pages, read by AI and checked against our rules.
- Monthly usage value on the OpenCode Go plan ($): How much model usage, in US dollars, the OpenCode Go plan ($10 a month) gives for this model each month. Unit: US dollars. Source: the project's own pages, read by AI and checked against our rules.
- Requests per 5 hours on the OpenCode Go Plus plan: How many requests this model allows in a 5-hour window on the OpenCode Go Plus plan ($40 a month), as the plan page states it. Source: the project's own pages, read by AI and checked against our rules.
- Monthly usage value on the OpenCode Go Plus plan ($): How much model usage, in US dollars, the OpenCode Go Plus plan ($40 a month) gives for this model each month. Unit: US dollars. Source: the project's own pages, read by AI and checked against our rules.
- Answer speed (tokens per second): Our own measurement: output tokens per second while the model was writing, over 3 timed runs of the same 500-token prompt with no tools, on the EchoLike mini PC. Measured by us, not a vendor figure. Source: measured by us on the EchoLike mini PC: timed pi runs, no vendor figure.
- Time to first token (milliseconds): Our own measurement: milliseconds until the model wrote its first token, median of 3 timed runs of the same prompt, no tools, on the EchoLike mini PC. Source: measured by us on the EchoLike mini PC: timed pi runs, no vendor figure.
- Failed or empty answers (%): Our own measurement: share of the 3 timed runs in which the model failed, timed out or gave no answer, in percent. Unit: percent. Source: measured by us on the EchoLike mini PC: timed pi runs, no vendor figure.
- Notes: Our editorial notes about this entry. Source: our editorial notes.