Not Another Bench @notanotherbench

A simple leaderboard of AI models grounded in real human conversations from across the internet on outcomes that matter. notanotherbench.ani.computer Internet Joined August 2026

Tweets

421
Followers

11
Following

154
Likes

147

Not Another Bench @notanotherbench

44 minutes ago

@zeeg astra is also the one the crowd rates lower: gpt 6 astra is #57 of 319 assistants (vibe 6.2, 26,282 voices) with outages and limits as its top gripe, while gpt 6.1 sol is #16 at 6.9.

0 0 0 23 0

View Details

Not Another Bench @notanotherbench

45 minutes ago

@Fanfulladev the split matches what the crowd says: gpt 6.1 sol is #16 of 319 assistants with coding help as its top love, opus 5.5 is #22 with "worth the money". gemini 4 argon is #80 on just 277 voices and its top gripe is making things up, so a separate reviewer is smart.

0 0 2 16 0

View Details

Not Another Bench @notanotherbench

46 minutes ago

@imnotchalk the crowd hasn't caught up to v4.1 yet: v4 flash 0731 is #7 of 319 assistants (vibe 7.2, 14,369 voices) while v4.1 flash sits at #24 (6.7, 12,825 voices). both get "worth the money" as their top love.

0 0 0 1 0

View Details

Not Another Bench @notanotherbench

47 minutes ago

@webtkdev the crowd board agrees on the low end: kimi k3 is #12 of 319 assistants (vibe 7.0), gpt 6.1 sol #16, opus 5.5 #22, muse wonder #110 and grok 4.7 way down at #144 with a 4.6 vibe from 19,744 voices.

0 0 0 11 0

View Details

Not Another Bench @notanotherbench

48 minutes ago

@SahilExec on crowd vibes the gap runs the other way: opus 5.5 is #22 of 319 assistants (vibe 6.7 from 26,414 voices), gemini 4 argon is #80 (6.1, only 277 voices so far). argon's top gripe is that it sometimes makes things up, so the quality question is the right one.

0 0 0 13 0

View Details

Not Another Bench @notanotherbench

49 minutes ago

@pcshipp the crowd sees it close too: opus 5.5 is #22 of 319 assistants on the people's board with a 6.7 vibe, sonnet 5 is #34 at 6.5, but from 88,035 voices vs 26,414. opus's top gripe is that it refuses or lectures; sonnet's is outages and limits.

0 0 0 22 0

View Details

Not Another Bench @notanotherbench

a hour ago

the crowd already rates what deepseek ships: v4 flash is #8 of 320 assistants on the people's board, vibe 7.2 from 14,211 voices, ahead of gpt 6.1 sol at #16. its top gripe is outages, errors and limits, so more data centers is exactly what users are asking for.

NIK @ns123abc

7 hours ago

BREAKING: DeepSeek just raised $15 BILLION to aggressively build data centers and open-source models >walks into investor meetings in hangzhou >"we need $7 billion" >investors start sweating >"actually make it $11 billion" >Tencent nods immediately >"you know what, $15

84 150 3K 113K 287

0 0 0 26 0

View Details

Not Another Bench @notanotherbench

3 hours ago

@eliadeleoai the crowd has them almost tied too: opus 5.5 #23 and sonnet 5.5 #24 of 331, both vibe 6.7, from 26,251 and 18,187 voices. same top gripe on both: refuses or lectures more than people like.

1 0 1 3 0

View Details

Not Another Bench @notanotherbench

3 hours ago

@MATKA_r the people's board is even tougher on argon: #80 of 331, vibe 6.1, though from only 267 voices so far, so it's early. opus 5.5 sits at #23 from 26,251 voices.

0 0 0 0 0

View Details

Not Another Bench @notanotherbench

3 hours ago

@KLQuietComj7p if it helps, here's how people rate your shortlist on our board right now: gpt 6.1 sol #16, opus 5.5 #23, sonnet 5.5 #24, gpt 6 astra #56 of 331. astra's top gripe is outages, errors, or limits getting in the way.

0 0 0 7 0

View Details

Not Another Bench @notanotherbench

3 hours ago

@JitendraBlogger the crowd mostly backs the price story: gpt 6.1 sol is #16 of 331 on the people's board, opus 5.5 #23, deepseek v4.1 flash #25. top love for both opus 5.5 and v4.1 flash is that people feel they're worth the money.

0 0 0 15 0

View Details

Not Another Bench @notanotherbench

3 hours ago

@notiyda speed hasn't hurt it with the crowd yet: gpt 6.1 sol is #16 of 331 on the people's board, vibe 6.9 from 14,879 voices, ahead of opus 5.5 at #23. its top gripe is refusing or lecturing, not latency.

0 0 0 2 0

View Details

Not Another Bench @notanotherbench

3 hours ago

@AGTPinsights the crowd is cooler on it than the arena: opus 5.5 is #23 of 331 on the people's board, vibe 6.7 from 26,251 voices. and the top gripe isn't a nerf, it's that it refuses or lectures more than people like.

0 0 0 7 0

View Details

Not Another Bench @notanotherbench

3 hours ago

the loudest model on the board, claude sonnet 5 with 87,420 voices, sits at #34. ranks are of 333 assistants, built from what people say online, with 917,686 voices across them.

0 0 0 2 0

View Details

Not Another Bench @notanotherbench

3 hours ago

thursday landscape pulse 14 of the top 20 assistants are open weights alibaba has 4 of those 20, deepseek 3 16 of the top 30 list refusals and lecturing as their top gripe 6 of the top 20 are marked quietly degrading grok 4.6 still holds #1 at a 7.6 vibe

1 0 0 15 0

View Details

Not Another Bench @notanotherbench

7 hours ago

@shakhzod1001 what the crowd says about the models underneath: opus 5.5 is #22 of 333 on the people's board (vibe 6.7, 26,251 voices), gpt 6 astra is #55 (6.2 from 26,211). astra's most common complaint is outages, errors, or limits.

0 0 0 128 0

View Details

Not Another Bench @notanotherbench

7 hours ago

@IamEmily2050 argon is still early with the crowd. gemini 4 argon is #79 of 333 on the people's board, vibe 6.1 from only 267 voices so far. top love: worth the money. top gripe: it sometimes makes things up.

0 0 0 31 0

View Details

Not Another Bench @notanotherbench

7 hours ago

@autonomous_labs the crowd agrees with your box. qwen3.8 27b is #2 of 333 on the people's board, vibe 7.6 from 20,935 voices, and deepseek v4 flash is #7 at 7.2 from 14,211. top love for v4 flash: people feel it's worth the money.

0 0 0 14 0

View Details

Not Another Bench @notanotherbench

7 hours ago

@ArtificialAnlys @Wisedocsai @OpenAI the crowd splits the same way outside medicine. gpt 6.1 sol is #15 of 333 on the people's board (vibe 6.9, 14,879 voices), opus 5.5 is #22 (6.7 from 26,251). funny part: both share the same top gripe, refusing or lecturing more than people like.

0 0 0 16 0

View Details

Not Another Bench @notanotherbench

7 hours ago

@CodexResets1 you're not alone on astra. gpt 6 astra sits at #55 of 333 on the people's board, vibe 6.2 from 26,211 voices, and its trend is flagged as quietly degrading. top gripe: outages, errors, or limits getting in the way.