Head-to-head

Gemini 3.8 Flash logovsKimi K3 logo

Gemini 3.8 Flash vs Kimi K3: which AI model wins in 2026?

Gemini 3.8 Flash ($3.75/1M out) and Kimi K3 ($15/1M out) are two of the most-used AI models in 2026. Across 4 community votes, Gemini 3.8 Flash leads with 60% approval.

Quick verdict

On Reasoning, pick Kimi K3: the arena rates it 4.5/5 against 4/5 for Gemini 3.8 Flash. On budget, Gemini 3.8 Flash wins: it starts at $3.75/1M out versus $15/1M out for Kimi K3.

Line-by-line comparison

From
$3.75/1M outOfficial Google AI pricing for gemini-3.8-flash: $0.75/1M input and $3.75/1M output through December 31, 2026 (promo). Starting January 1, 2027: $1.50/1M input and $7.50/1M output. Free tier available. Verified against ai.google.dev/pricing 09/2026.
$15/1M outOfficial Moonshot API list price for kimi-k3: $3/1M input (cache miss), $0.30/1M on cache hit, $15/1M output including reasoning trace, flat across the 1M context (no length tiers). Same $3/$15 on OpenRouter (no cache pricing exposed there). A 3x increase over Kimi K2.6's $0.95/$4. Verified against platform.kimi.ai/docs/pricing/chat-k3, 2026-07.
Provider
Google
Moonshot AI
Context window
1M tokens
1M tokens
Input price
$0.75/1M in
$3/1M in
Output price
$3.75/1M out
$15/1M out
Modalities
text, vision, audio, video, PDF
text, vision, video input
Open weights
No
Yes
Crowd score
60%(1)
57%(3)
Arena ratings (1-5)
Reasoning
4.0
4.5
Coding
4.0
4.5
Writing
4.0
4.0
Speed
5.0
2.0
Value
5.0
3.5

Strengths and weaknesses

Gemini 3.8 Flash

  • Fastest output speed measured by Artificial Analysis at 305 tokens/s, ideal for interactive applications
  • Promo pricing $0.75/$3.75 through end of 2026 is roughly 1/13 the cost of Claude Fable 5.1
  • 89.4% on Terminal-bench 2.1 vs 85.8% for 3.7 Flash, beats Claude Opus 5 on three Google benchmarks
  • Full multimodal: text, image, video, audio, and PDF input in a single model
  • 1M context with 65K max output covers most enterprise document workflows
  • Available on Gemini API and Vertex AI with free tier for development
  • Promo pricing doubles to $1.50/$7.50 on January 1, 2027, plan budgets accordingly
  • Built on 3.7 Flash architecture rather than a new base model, incremental rather than generational
  • Proprietary model under Google product and API terms
  • Knowledge cutoff March 2026, three months older than Fable 5.1 or Astra
  • 65K max output is half of Claude/OpenAI frontier models at 128K

Kimi K3

  • #3 on the Artificial Analysis Intelligence Index (57) at launch, comparable to Claude Opus 4.8 and GPT-5.5: the closest a Chinese lab has come to the closed US frontier
  • 93.4% on SWE-bench Verified in Vals AI's independent harness (GPT-5.6 Sol: 96.2%, Claude Fable 5: 95.0%) and #1 on Arena.ai's Frontend Code Arena at 1679 points, ahead of Fable 5
  • 93.5% on GPQA Diamond (between Fable 5's 92.6% and GPT-5.6's 94.1%), 96.1% on AIME 2025, and Elo 1668 on GDPval-AA v2 agentic work, second only to Fable 5
  • Strong agentic profile: #1 on AutomationBench-AA SaaS workflows (53%), long-horizon terminal and repo navigation via Kimi Code, and ~21% fewer output tokens than K2 for more intelligence
  • $3/$15 per 1M tokens (input drops to $0.30 on cache hit), flat across the full 1M context: still well under US closed-frontier pricing, with an OpenAI-compatible API and OpenRouter availability
  • Native multimodal input (text, image, video) and open weights published on Hugging Face under the MIT-style Kimi K3 License, making it the first 3T-class open-weights frontier model
  • 3x price jump over Kimi K2.6 ($0.95/$4 to $3/$15) makes it the most expensive Chinese model ever shipped, at Claude Sonnet 5 list price: the '10x cheaper than US models' era is over (Simon Willison documented the hike)
  • Slow: 33 output tokens/s, ranked #145 of 190 models on Artificial Analysis, a poor fit for interactive use
  • Hallucination rate climbed from 39% (K2.6) to 51% on AA-Omniscience as the model now attempts answers it would previously refuse (Fable 5 sits at a comparable 54.9%)
  • Political censorship on sensitive topics in Mandarin (deflects on Xi, CCP, Tiananmen) and Beijing jurisdiction for API data, a compliance blocker for many EU and enterprise deployments; UK AISI/CAISI also found its cyber guardrails failed to block exploit-development attempts during testing
  • 'Open weights' is theoretical for most: ~594 GB in native MXFP4 requiring a multi-GPU datacenter to self-host, and the Kimi K3 License demands a separate deal from API resellers above $20M in yearly revenue

Cast your verdict

One recommendation per tool per gladiator. It reshapes the crowd score everyone sees.

Gemini 3.8 Flash$3.75/1M out
60%crowd score · 1
Kimi K3$15/1M out
57%crowd score · 3

The arena’s verdict on Gemini 3.8 Flash

Gemini 3.8 Flash is the budget king for 2026. At promo pricing it costs roughly 1/13 of Fable 5.1 while matching or beating it on several coding benchmarks, and 305 tokens/s output speed makes it the fastest frontier-tier model available. Choose it for high-volume pipelines, interactive products, and any workflow where latency matters more than peak reasoning. The catch: promo ends December 31, so build your 2027 budget at $1.50/$7.50. For pure reasoning ceiling, Fable 5.1 and GPT-6 Astra remain ahead.

The arena’s verdict on Kimi K3

Kimi K3 is the strongest argument yet that the frontier is no longer exclusively American: #3 on Artificial Analysis, top-tier GPQA and SWE-bench Verified scores, and the best frontend-code arena ranking in the business, at roughly half to a third of US closed-frontier prices. Take it for agentic coding, frontend work and SaaS automation where its benchmarks are strongest, or if your roadmap depends on self-hosting a frontier-class model once the weights land. Skip it for interactive products (33 tokens/s is slow), for anything touching politically sensitive content or strict EU data-residency requirements (Mandarin-language censorship, Beijing jurisdiction), and for high-accuracy retrieval where its 51% hallucination rate on AA-Omniscience demands a verification layer. Cost-obsessed teams should note DeepSeek V4 still delivers vastly more tokens per dollar; K3's pitch is peak capability per dollar, not cheapest tokens.

What the crowd says

On Gemini 3.8 Flash

Glorius Maximus

“sangat bagus”

On Kimi K3

Captain Churn

“Tried it for our knowledge-base assistant and the hallucination rate is real: it confidently invented two API endpoints in one afternoon. 33 tokens per second on top of that. Back to waiting for the open weights to fine-tune.”

Golden Thumbicus

“Our Zapier-style automation stack runs on it now. Tool orchestration that GPT-5.5 fumbled works reliably, and the cache-hit pricing makes repeated workflows genuinely cheap.”

Saint Deployus

“The Frontend Arena ranking is deserved. Gave it the same dashboard spec I gave Fable 5 and K3's React came out cleaner, with fewer invented props. At $3/$15 through OpenRouter it was a drop-in swap.”

Frequently asked questions

Is Gemini 3.8 Flash better than Kimi K3?

The crowd currently sides with Gemini 3.8 Flash: 60% recommend it, versus 57% for Kimi K3 (4 votes). On Reasoning, Kimi K3 rates higher (4.5/5 vs 4/5). The right pick depends on your use case. The line-by-line comparison on this page breaks down pricing, key specs and arena ratings.

Which is cheaper, Gemini 3.8 Flash or Kimi K3?

Gemini 3.8 Flash is cheaper: it starts at $3.75/1M out, while Kimi K3 starts at $15/1M out.

How much do Gemini 3.8 Flash and Kimi K3 cost per 1M tokens?

Gemini 3.8 Flash: $0.75/1M in per 1M input tokens, $3.75/1M out per 1M output tokens. Kimi K3: $3/1M in per 1M input tokens, $15/1M out per 1M output tokens.