Real head-to-head · same prompt, one shot

Fugu Mini vs MiniMax M3

Fugu's fast mini variant — single model, no panel, ~3 min per build. vs 1M-context frontier model at $0.30/M tokens — cheapest big-context model on the bench.

Head-to-head verdict: Fugu Mini wins 1–0.

Fugu Mini · contextSakana subscription · same key as Ultra
MiniMax M3 · context1M tokens
Fugu Mini · priceSame Sakana subscription pool as Fugu Ultra
MiniMax M3 · price$0.30 / 1M input tokens, $1.50 / 1M output
Fugu Mini · vendorSakana AI
MiniMax M3 · vendorMiniMax

What I tested — same prompt, two models

I run the same fixed prompt set through every new model the day it drops — same string, one shot, single HTML file out — and I score the result 0–10 on whether it ran, how close it hit the brief, and how good it looked. Below is what came out when I gave the exact same prompts to Fugu Mini and MiniMax M3, side by side, on 13 shared tasks inside the Agent Operating System.

Both models were given identical prompts inside the Agent Operating System — no help, no iteration, no "best of N" tricks. I run each prompt once, save the HTML file the model produces, and score it 0–10 on whether it ran, how close it hit the brief, and how good it looked. The scoring is mine. The verdicts below are pulled from my source comparison guides at agentos.guide where I publish every score and the reasoning behind it.

Fugu Mini · Dispatched from Agent OS as the fast Sakana lane. Bench scored by Claude judge against the same 42 prompts.

MiniMax M3 · Bench prompts dispatched via OpenRouter. Scored by Claude judge against the same 42 prompts every other model ran.

Side-by-side on 42 shared tasks

Click any cell to play that model's actual one-shot attempt. Medals are derived from my 0–10 scores per task (highest = 🥇, second = 🥈, third = 🥉).

Task ↓
Fugu Mini
MiniMax M3
Game
Fugu Mini on Arcade
🥉MiniMax M3 on Arcade
Game
Fugu Mini on Doom
MiniMax M3 on Doom
Game
Fugu Mini on Outrun
MiniMax M3 on Outrun
Page
Fugu Mini on Landing
MiniMax M3 on Landing
Sim
Fugu Mini on Blackhole
MiniMax M3 on Blackhole
Sim
Fugu Mini on Fluid
🥉MiniMax M3 on Fluid
Sim
Fugu Mini on Fractal
🥈MiniMax M3 on Fractal
Sim
Fugu Mini on Galaxy
MiniMax M3 on Galaxy
Sim
Fugu Mini on Orbit
🥈MiniMax M3 on Orbit
Sim
MiniMax M3 on Solar
Visual
Fugu Mini on Aurora
🥇MiniMax M3 on Aurora
Visual
Fugu Mini on Matrix
🥇MiniMax M3 on Matrix
Visual
Fugu Mini on Terrain
🥈MiniMax M3 on Terrain
Game
— not attempted —
🥈MiniMax M3 on Crypt
Game
— not attempted —
🥈MiniMax M3 on Dogfight
— not attempted —
🥈MiniMax M3 on Dragonflight
— not attempted —
🥇MiniMax M3 on Dragonrealm
Game
— not attempted —
🥉MiniMax M3 on Game
— not attempted —
🥈MiniMax M3 on Neonblaster
Game
— not attempted —
MiniMax M3 on Neoncity
Game
— not attempted —
🥈MiniMax M3 on Neonracer
— not attempted —
🥈MiniMax M3 on Nordiccrypt
Game
— not attempted —
🥉MiniMax M3 on Pool
Game
— not attempted —
🥇MiniMax M3 on Racing

Where Fugu Mini beat MiniMax M3

The tasks where I gave Fugu Mini a higher 0–10 score on the same prompt — with the actual commentary from my source guides.

Solar Sim
Fugu Mini 8.0 · MiniMax M3 7.5 (+0.5)

What I saw: Single-model Fugu Mini shipped a complete 12KB solar system in 3.5 min — full </html>, animation loop, Saturn rings, drag/orbit. Cleaner than Ultra's panel attempts which mostly timed out.

Strengths & weaknesses I logged

Fugu Mini

Strengths

  • Zero panel orchestration — much lower latency than Ultra
  • Same Sakana subscription, no extra cost
  • Doesn't time out on heavy game/3D prompts where Ultra stalls

Trade-offs

  • Single model only — no ensemble verdict
  • Newer than Ultra — less calibration / verification

MiniMax M3

Strengths

  • 1M token context — full repo / full deep-research corpus fits in one call
  • $0.30/M input is roughly 1/30th of Opus 4.8 — built for high-volume agent loops
  • Solid one-shot HTML output — clean structure on game and visual prompts

Trade-offs

  • Less polished than Fusion's panel-ensembled output on the toughest deep builds
  • Newer model — less community calibration vs Fable 5 / Opus 4.8

Pricing & context — the spec sheet

Spec Fugu Mini MiniMax M3
VendorSakana AIMiniMax
Context windowSingle-model variant of Sakana's Fugu — no panel orchestration. Same API endpoint, much faster per call.1,048,576-token context — matches GLM-5.2 and Fable 5
PriceSame Sakana subscription pool as Fugu Ultra$0.30 / 1M input tokens, $1.50 / 1M output
Pricing detailThe non-Ultra `fugu` model on Sakana's API. Sakana describes it as 'Fast mini model optimized for low latency yet high quality responses.' Crucially: zero orchestration tokens per call (vs Ultra's panel of thousands). Returns in ~3 min instead of 6-15 min and doesn't time out on heavy prompts.MiniMax M3 is the cheapest 1M-context frontier model on the bench — roughly 1/200th the per-call cost of OpenRouter Fusion and 1/30th of Claude Opus 4.8. Designed for high-volume agent workloads where context length matters but per-call budget is tight.
Release2026-06-152026-06-18
Bench coverage1/13 scored · avg 8.00/1042/42 scored · avg 7.96/10

The verdict — which should you pick?

Across 1 scored shared tasks, Fugu Mini averaged 8.00/10, beating MiniMax M3's 7.50/10 by 0.50 points. Pick Fugu Mini when the build has to ship on the first prompt and you can afford the trade-offs in the comparison below.

If you only run one of these inside your stack, the head-to-head average above is the call. If you can run both, my honest play is to wire Fugu Mini and MiniMax M3 both into the Agent Operating System and dispatch each from the kanban by task type — agent loops where latency matters more than panel consensus → Fugu Mini, high-volume agent workflows where per-call cost dominates → MiniMax M3. That's the same setup I run for the 3,600+ founders inside the AI Profit Boardroom.

FAQ — Fugu Mini vs MiniMax M3

Which is better, Fugu Mini or MiniMax M3?

On Goldie Bench, Fugu Mini averages 8.00/10 across the shared tasks, with 0 gold, 0 silver, 0 bronze overall. MiniMax M3 averages 7.50/10, with 7 gold, 14 silver, 10 bronze. Fugu Mini wins the head-to-head 1–0.

How much does Fugu Mini cost vs MiniMax M3?

Fugu Mini: The non-Ultra `fugu` model on Sakana's API. Sakana describes it as 'Fast mini model optimized for low latency yet high quality responses.' Crucially: zero orchestration tokens per call (vs Ultra's panel of thousands). Returns in ~3 min instead of 6-15 min and doesn't time out on heavy prompts. MiniMax M3: MiniMax M3 is the cheapest 1M-context frontier model on the bench — roughly 1/200th the per-call cost of OpenRouter Fusion and 1/30th of Claude Opus 4.8. Designed for high-volume agent workloads where context length matters but per-call budget is tight.

What's the context window for Fugu Mini vs MiniMax M3?

Fugu Mini has a Single-model variant of Sakana's Fugu — no panel orchestration. Same API endpoint, much faster per call. context window. MiniMax M3 has a 1,048,576-token context — matches GLM-5.2 and Fable 5 context window.

When should I pick Fugu Mini over MiniMax M3?

Pick Fugu Mini for: Agent loops where latency matters more than panel consensus; Quick first-drafts you'll refine downstream; Filling out a bench when Ultra is timing out. The trade-off is the weaknesses we logged on the bench: Single model only — no ensemble verdict; Newer than Ultra — less calibration / verification.

When should I pick MiniMax M3 over Fugu Mini?

Pick MiniMax M3 for: High-volume agent workflows where per-call cost dominates; 1M-context tasks (whole-repo refactors, deep-research synthesis); Drop-in cheaper alternative to GLM-5.2 with comparable 1M context. The trade-off is the weaknesses we logged on the bench: Less polished than Fusion's panel-ensembled output on the toughest deep builds; Newer model — less community calibration vs Fable 5 / Opus 4.8.

How does Goldie Bench score Fugu Mini vs MiniMax M3?

Every demo on this page was built by Julian Goldie inside the Agent Operating System — same fixed prompt for both models, one shot, single HTML file out. Each result gets a 0–10 score on whether it ran, how close it hit the brief, and how good it looked. The highest score on each task gets gold; second gets silver; third gets bronze. See methodology for full provenance.

The same stack Julian uses

Run this stack yourself.

Every demo on this bench was built inside the Agent Operating System — one prompt, one shot, single HTML file out. The Agent OS, the prompts, the templates, the weekly walkthroughs and 3,600+ founders shipping with it every day all live inside the AI Profit Boardroom.

3,600+founders
258documented wins
38countries
$100k+/mocommunity MRR