
Kimi K2.7 vs Hy3
The heavy lifter — frontier coder at flat-rate. vs Tencent's open-weights coder — Apache-2.0, cheap, beats GLM-5.1 on frontend in Tencent's blind eval.
Head-to-head verdict: Kimi K2.7 wins 4–1 with 1 tie.
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 Kimi K2.7 and Hy3, side by side, on 7 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.
Kimi K2.7 · Wired into the Agent OS as the heavy-lifter for game/sim prototypes and Kanban-dispatched code work. Mode toggled per task: Quality for one-shot games, Fast for short bursts.
Hy3 · Wired into the Agent OS as the 'Hy3 Coder' tab (chat + live preview + workspace) via OpenRouter. Bench built one-shot on the same prompts as the field; weak builds iterated by Hy3 itself (the model fixes its own builds).
Side-by-side on 47 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 = 🥉).
Where Kimi K2.7 beat Hy3
The tasks where I gave Kimi K2.7 a higher 0–10 score on the same prompt — with the actual commentary from my source guides.
What I saw: All three are real, playable shooters. Opus drops you in a corridor with an imp dead ahead — gun, crosshair and HUD framed like a screenshot. Kimi matches it: a monster down a textured hall, health, ammo, minimap. GLM ships a gorgeous 'HAZARD PROTOCOL' title screen with a working…
What I saw: 23KB · plays clean · three, webgl
What I saw: 35KB · plays clean · three (re-rolled)
What I saw: 40KB · plays clean · three
Where Hy3 beat Kimi K2.7
The tasks where I gave Hy3 a higher 0–10 score on the same prompt — with the actual commentary from my source guides.
What I saw: Clean multi-scene pipeline with animated count-up stats, gold/cyan cinematic palette, particle field and progress bar render correctly; but the stats appear left-clustered and off-center with only two of four visible mid-animation, feeling sparse rather than composed, keeping it …
Strengths & weaknesses I logged
Kimi K2.7
Strengths
- Best-of-three on interactive games — raycaster, DOOM, monster AI
- Three speed modes (Fast / No-Think / Quality) you can swap per task
- Flat-rate plan eliminates the per-token meter, so iteration is free
Trade-offs
- Plays plainest on abstract visual prompts — synthwave grids, fluid sims, aurora — where GLM and Opus add more flair
- Bronze average on the Goldie Bench bench despite the gold-medal games — its visual builds are accurate but understated
Hy3
Strengths
- Apache-2.0 open weights — self-host free, no lock-in
- Tencent's 270-expert blind eval: 2.67/4 vs GLM-5.1's 2.51, strongest on frontend / data / CI-CD
- Hallucination rate cut 12.5% → 5.4%; stable tool-calls across scaffoldings (<4% SWE-Bench variance)
Trade-offs
- Slow upstream on OpenRouter (30-90s per build) — fine for one-shots, sluggish for tight loops
- One-shot game builds can under-render (flat raycaster walls, unlit 3D) without an iterate pass
Pricing & context — the spec sheet
| Spec | Kimi K2.7 | Hy3 |
|---|---|---|
| Vendor | Moonshot AI | Tencent Hunyuan |
| Context window | 256,000 tokens | 262,144-token context window. Open weights (Apache-2.0) on HuggingFace / ModelScope / GitHub; benched here via OpenRouter. |
| Price | Flat plan (no per-token bill) | $0.14 / 1M input · $0.58 / 1M output |
| Pricing detail | Available on Moonshot's flat-rate subscription plan — no per-token billing for individual builders. The plan covers all three speed modes (Fast, No-Think, Quality). Vendor: Moonshot AI (moonshot.ai), based in Beijing. | Tencent Hunyuan 3 — open-weights under Apache-2.0, so free to self-host. On OpenRouter it is one of the cheapest capable coders: ~$0.14/M in, $0.58/M out (1 RMB / 4 RMB). Upstream can be slow (30-90s to first token), but per-token cost is negligible. |
| Release | 2026-06 | 2026-07-06 |
| Bench coverage | 25/47 scored · avg 7.46/10 | 7/7 scored · avg 7.13/10 |
The verdict — which should you pick?
Across 6 scored shared tasks, Kimi K2.7 averaged 7.75/10, beating Hy3's 7.12/10 by 0.63 points. Pick Kimi K2.7 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 Kimi K2.7 and Hy3 both into the Agent Operating System and dispatch each from the kanban by task type — interactive game prototypes you want shippable on the first prompt → Kimi K2.7, cost-sensitive coding + frontend design where open weights matter → Hy3. That's the same setup I run for the 3,600+ founders inside the AI Profit Boardroom.
FAQ — Kimi K2.7 vs Hy3
Which is better, Kimi K2.7 or Hy3?
On Goldie Bench, Kimi K2.7 averages 7.75/10 across the shared tasks, with 2 gold, 3 silver, 0 bronze overall. Hy3 averages 7.12/10, with 0 gold, 1 silver, 0 bronze. Kimi K2.7 wins the head-to-head 4–1.
How much does Kimi K2.7 cost vs Hy3?
Kimi K2.7: Available on Moonshot's flat-rate subscription plan — no per-token billing for individual builders. The plan covers all three speed modes (Fast, No-Think, Quality). Vendor: Moonshot AI (moonshot.ai), based in Beijing. Hy3: Tencent Hunyuan 3 — open-weights under Apache-2.0, so free to self-host. On OpenRouter it is one of the cheapest capable coders: ~$0.14/M in, $0.58/M out (1 RMB / 4 RMB). Upstream can be slow (30-90s to first token), but per-token cost is negligible.
What's the context window for Kimi K2.7 vs Hy3?
Kimi K2.7 has a 256,000 tokens context window. Hy3 has a 262,144-token context window. Open weights (Apache-2.0) on HuggingFace / ModelScope / GitHub; benched here via OpenRouter. context window.
When should I pick Kimi K2.7 over Hy3?
Pick Kimi K2.7 for: Interactive game prototypes you want shippable on the first prompt; High-iteration agent loops where per-token cost would dominate; Long-context refactors using the 256K window inside Agent OS. The trade-off is the weaknesses we logged on the bench: Plays plainest on abstract visual prompts — synthwave grids, fluid sims, aurora — where GLM and Opus add more flair; Bronze average on the {{SITE_NAME}} bench despite the gold-medal games — its visual builds are accurate but understated.
When should I pick Hy3 over Kimi K2.7?
Pick Hy3 for: Cost-sensitive coding + frontend design where open weights matter; Self-hosters who want an Apache-2.0 model they fully own; Anyone wiring a cheap capable coder into a live build panel (Agent OS Hy3 Coder tab). The trade-off is the weaknesses we logged on the bench: Slow upstream on OpenRouter (30-90s per build) — fine for one-shots, sluggish for tight loops; One-shot game builds can under-render (flat raycaster walls, unlit 3D) without an iterate pass.
How does Goldie Bench score Kimi K2.7 vs Hy3?
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.
Related comparisons
Other head-to-heads using the same scoring system:
Kimi K2.7 vs Fusion Hy3 vs Fusion Kimi K2.7 vs Hermes MoA Hy3 vs Hermes MoA Kimi K2.7 vs Claude Fable 5 Hy3 vs Claude Fable 5 Kimi K2.7 vs Grok Hy3 vs GrokFull model pages: Kimi K2.7 · Hy3 · back to the leaderboard
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.





























