ai leaks and takes | deep in love with claude
- Tweets 2,736
- Following 124
- Followers 309
- Likes 8,898
the ultimate benchmark test has arrived.
gpt-6 luna, deepseek, gemini and claude
same prompt, who wins round 1???
7
36
1,574
got grok plus a week ago. fast, very good at design, honestly impressive but it only has a weekly limit, no 5 hour window
feels good until you burn it all in one day and you're locked out till friday
@elonmusk a daily reset would fix this
5
10
528
minimax m3.1 flash looks great, until you put it next to deepseek v4.1 flash 1017
deepseek is cheaper. 0.058 vs 0.15 per million input tokens, and cache is 5x cheaper too
it also wins at coding on minimax's own chart.
77 vs 73.7 on deepswe and deepseek weights are already out, minimax drops theirs oct 23
nice launch, but deepseek still wins this one
3
231
honest question. how much do you spend per month to test all these new models? subs, api, hardware, all of it
3
5
288
for 0.15 and 0.60 per million tokens this is amazing for a flash model. but they compare against opus 5, not 5.5, and opus 5 still wins most of the coding rows
MiniMax M3.1 Flash: Space Performance + Bunny Speed
🚀🐰 Yes, Space Bunny - which has topped global leaderboards with 200T+ tokens - was an earlier version of M3.1 Flash.
💰Redefining the performance economics of agentic AI. M3.1 Flash achieves frontier-level intelligence across key tasks such as Cowork (Automation-Bench: 62%; ALE: 30%), Coding (DeepSWE: 74%) and Visual Reasoning (BabyVision: 91%), at 1/3 the cost of M3 and 3x the speed.
API pricing (flat rate): $0.15 / $0.60 per 1M input/output tokens, with cache reads at $0.015 / 1M tokens.
Model weights will arrive on Oct 23.
API: platform.minimax.io
M Plan: minimax.io/m-plan
MiniMax Code for desktop: agent.minimax.io/download
1
6
318
tomorrow is tuesday. half the timeline says fable 5.5 drops oct 27 instead, which is also a tuesday. so the pattern lives either way
anthropic loves tuesdays
fable 5: june 9
fable 5.1: sept 1
opus 5.5: sept 22
next tuesday is oct 20
if fable 5.5 doesn't land then, anthropic breaks its own pattern
2
13
1,407
