K
Kimi K3.1 Pro
2.8 trillion parameters and first on Terminal-Bench 4.0, but output runs at just 28 tok/s
Intelligence v5
46.0
Leaderboard rank
#9
Input / output
$2.00/$10.00
Blended 3:1
$4.00
Output speed
28 tok/s
Value
12 pts/$
Capability profile Normalised to the field maximum
Specifications
Vendor
Moonshot AI
Released
2026-10-23
Weights
Open (Kimi K3 License)
Context
1,000,000 tokens
Max output
Undisclosed
Parameters
2800B total / 104B active
Input price
$2.00 / 1M
Output price
$10.00 / 1M
Cache hit
$0.20 / 1M
Output speed
28 tok/s
Intelligence
46.0（v5）
Benchmark scores
How it is measured →Ranked 9 · bars are scaled to the field maximum for that benchmark.
| Benchmarks | Score | Vs. best in field | Gap |
|---|---|---|---|
| AA Intelligence Index v5 Intelligence index | 46.0 | 77% | −13.4 pts |
| AA Intelligence Index v4.3.2 Intelligence index (legacy) | 55.1 | 96% | −2.5 pts |
| GPQA Diamond | — | — | not indexed |
| Terminal-Bench 4.0 | — | — | not indexed |
| LiveBench Design Document generation (DOCX/PPT/PDF) | 77.8 | 90% | −9.1 pts |
| Vals Index | — | — | not indexed |
Assessment
A post-trained version of Kimi K3 with the same parameter count (2.8T total / 104B active) and 1M context, at two thirds of K3's price. First on Terminal-Bench 4.0 at 80.90% (legacy Vals scale), but its 28 tok/s output is the slowest of any top model, so long outputs take noticeably longer.
At the same price
The four models closest on blended price, side by side.
| Model | v5 | Blended | tok/s | Score gap |
|---|---|---|---|---|
| Claude Sonnet 5.5Anthropic | 54.2 | $4.00 | 135 | +8.2 |
| GPT-6.1 SolOpenAI | 54.5 | $4.00 | 77 | +8.5 |
| Gemini 4 CarbonGoogle | 59.4 | $4.00 | 87 | +13.4 |
| Gemini 4 ArgonGoogle | 50.4 | $4.00 | 80 | +4.4 |
Compare side by side
Line it up against other models on price, speed and every benchmark.
The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.
Open the comparison tool → Back to leaderboard