0
Fork 0
mirror of https://github.com/mem0ai/mem0.git synced 2026-11-17 19:02:25 +00:00

Compare commits

...

19 commits

Author SHA1 Message Date
AGAMPANDEYY
ba65f0bc0a docs(platform): BYOK chart: the green area under $10 per 1,000 add requests 2026-10-17 19:38:22 +05:30
AGAMPANDEYY
80bb0eb580 docs(platform): BYOK model guide: the new picks, the cheaper models and the new chart
Best value is DeepSeek V4.1 Flash, Best under $1 is Amazon Nova Micro, Best quality stays Claude Opus 5.5. The table
adds Amazon Nova Lite, Amazon Nova Micro, Gemma 3 4B, GPT-4.1 nano and Mistral Nemo. The chart is the redesigned one
from the blog, and the cost line no longer says list prices.
2026-10-17 17:17:58 +05:30
AGAMPANDEYY
cd07f73b26 docs(platform): BYOK model guide: the full evaluation no longer breaks accuracy down by question type 2026-10-16 23:49:15 +05:30
AGAMPANDEYY
efd16607db docs(platform): BYOK model guide: end the description on "your use case" 2026-10-16 22:47:46 +05:30
AGAMPANDEYY
4c927b9910 docs(platform): BYOK model guide: say what the page helps you do
Vinay's review: the description now starts with the reader's task, comparing
models and picking one.
2026-10-16 22:07:25 +05:30
AGAMPANDEYY
82da836bc5 docs(platform): BYOK model guide: add requests and memory extraction, as the pricing page says
Cost is per 1,000 add requests (the unit on the pricing page) and the model is
described as the one that extracts memories. The rate-limit note is gone and the
images carry the same unit.
2026-10-16 22:07:25 +05:30
AGAMPANDEYY
8e0cbc85f7 docs(platform): BYOK page: a heading and lead line for the picks and for the results 2026-10-16 22:07:25 +05:30
AGAMPANDEYY
6249c30f2d docs(platform): BYOK page: the three picks as the blog's cards 2026-10-16 22:07:25 +05:30
AGAMPANDEYY
85aa72696f docs(platform): BYOK page: full-width pick cards and a results table that fits the page 2026-10-16 22:07:24 +05:30
AGAMPANDEYY
95421f4cb9 docs(platform): BYOK page: plain How we tested, owner-only setup, MiMo as the OpenRouter example 2026-10-16 22:07:24 +05:30
AGAMPANDEYY
44344f0df3 docs(platform): BYOK page: read-more link to the full evaluation at the top 2026-10-16 22:07:24 +05:30
AGAMPANDEYY
8068718f3f docs(platform): BYOK page: no count of the tied models, default model at 92.5% 2026-10-16 22:07:24 +05:30
AGAMPANDEYY
4302b636de docs(platform): BYOK page: no internal steps in How we tested 2026-10-16 22:07:24 +05:30
AGAMPANDEYY
b12b9770d9 docs(platform): BYOK page links to the post's new URL 2026-10-16 22:07:24 +05:30
AGAMPANDEYY
616a454242 docs(platform): BYOK chart: each model's label on one line 2026-10-16 22:07:24 +05:30
AGAMPANDEYY
91a2f981c7 docs(platform): BYOK chart: best-for-the-price line back
On the averaged numbers the line runs MiMo-V2.6-Flash, Claude Sonnet 5.5 and Claude Opus 5.5.
2026-10-16 22:07:24 +05:30
AGAMPANDEYY
7e4e5c44de docs(platform): BYOK accuracy as the average of two runs
Each model's accuracy is now the average of its first run and a second that gives every missed question one more try. The free tier shows its published 92.5%. Seven models are tied at the top, between 92.6% and 93.5%.
2026-10-16 22:07:24 +05:30
AGAMPANDEYY
e160e8c1cd docs(platform): BYOK picks: best value MiMo-V2.6-Flash, best on OpenAI GPT-6 Luna
MiMo is the cheapest model tested and tied at the top, so it takes best value and the cheapest card goes. The chart drops its best-for-the-price line.
2026-10-16 22:07:24 +05:30
AGAMPANDEYY
4e1d8fabcd docs(platform): add the BYOK model guide
Accuracy, cost per 1,000 adds and p95 latency for the 13 models we evaluated on LoCoMo, the three picks, the chart, and the steps to add a key under Settings -> BYOK. Links to the full evaluation on the blog.
2026-10-16 22:07:24 +05:30
5 changed files with 73 additions and 1 deletions

Binary file not shown.

After

Width:  |  Height:  |  Size: 441 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 75 KiB