Kilo intelligence
Benchmark signal / live operational data
AI model
performance
report
A concise view of coding completion, cost efficiency, model popularity, and inference provider reliability across the Kilo ecosystem.
- Published
- November 5, 2026
- Benchmark
- Terminal Bench 2.0
- Inference window
- Latest weekly snapshot
01 / Capability and cost
KiloBench leaders
Models are ranked by completion rate on Terminal Bench 2.0. Cost per attempt is the observed average for a complete benchmark attempt, including unsuccessful attempts.
Top completion
79.3%
GPT-6 Astra ($$$$)
Top model cost
$107.29
Average cost per attempt
Lowest top-10 cost
$2.58
Among ranked models shown
Average completion
74.5%
Mean across the top 10
| # | Model | Provider | Completion rate | Cost / attempt |
|---|---|---|---|---|
| 1 | GPT-6 Astra ($$$$) 5 attempts | openai | 79.3% | $107.29 |
| 2 | GPT-5.6 Sol 5 attempts | openai | 76.2% | $87.41 |
| 3 | Claude Fable 5.1 ($$$$) 5 attempts | anthropic | 76.2% | $91.41 |
| 4 | DeepSeek V4.1 Flash 5 attempts | deepseek | 75.3% | $2.58 |
| 5 | Gemini 3.8 Flash (50% off) 5 attempts | 75.3% | $112.27 | |
| 6 | GPT-5.5 5 attempts | openai | 74.2% | $72.63 |
| 7 | Grok 4.6 5 attempts | x-ai | 73.0% | $33.83 |
| 8 | Kimi K3 5 attempts | moonshotai | 72.8% | $48.38 |
| 9 | Claude Opus 5 5 attempts | anthropic | 71.5% | $113.54 |
| 10 | Claude Fable 5 ($$$$) 5 attempts | anthropic | 71.0% | $87.52 |
How to read this page. Completion rate is the KiloBench score: the share of benchmark tasks completed successfully. Published results require at least five attempts and an observed attempt cost; inactive and stealth models are excluded. Rankings prioritize higher completion, then lower cost when scores tie.
02 / Real-world demand
Inference market view
The 20 most-used models are ranked by token volume from the latest Kilo Gateway weekly snapshot. Primary-provider metrics describe the provider carrying the largest traffic share for each model.
Observed volume
6.5T
Tokens across published records
Models ranked
20
Ordered by usage volume
Providers observed
47
Distinct inference providers
Most-used model
#01
nemotron-3.5-super-vl:free
| # | Model | Primary provider | Usage volume | Cost / request | Success rate |
|---|---|---|---|---|---|
| 1 | nemotron-3.5-super-vl:free 1 provider | NVIDIA | 1.7T | $0 | 95.45% |
| 2 | nemotron-3-ultra-550b-a55b:free 1 provider | NVIDIA | 1.2T | $0 | 92.90% |
| 3 | beam-free 1 provider | Reflection | 1.1T | $0 | 93.99% |
| 4 | glyph-cluster 1 provider | Stealth | 567.6B | $0 | 96.24% |
| 5 | dots-3-note-preview:free 1 provider | Atlas Cloud | 487.9B | $0 | 87.96% |
| 6 | deepseek-v4.1-flash-1017 8 providers | Novita AI | 258.9B | $0.0038 | 99.82% |
| 7 | laguna-s-2.1:free 1 provider | Poolside | 151.4B | $0 | 96.57% |
| 8 | deepseek-flash-latest 5 providers | Novita AI | 125.1B | $0.0025 | 88.02% |
| 9 | nemotron-3.5-lightning:free 1 provider | NVIDIA | 123.6B | $0 | 82.04% |
| 10 | deepseek-v4.1-flash 3 providers | Novita AI | 104.6B | $0.0025 | 99.49% |
Primary provider. The provider with the highest token volume for a model. Success rate equals 100% minus observed error rate. Costs are average observed request costs during the reporting period; zero indicates free traffic in the source data.
02 / Real-world demand
Inference market view / continued
The 20 most-used models are ranked by token volume from the latest Kilo Gateway weekly snapshot. Primary-provider metrics describe the provider carrying the largest traffic share for each model.
| # | Model | Primary provider | Usage volume | Cost / request | Success rate |
|---|---|---|---|---|---|
| 11 | openrouter/free 8 providers | NVIDIA | 95.0B | $0 | 87.46% |
| 12 | gpt-6.1-sol 3 providers | Openai Chatgpt | 82.8B | $0 | 99.46% |
| 13 | claude-opus-5.5 4 providers | 67.8B | $0.16 | 99.27% | |
| 14 | gpt-6-luna 3 providers | OpenAI | 56.5B | $0.0033 | 99.43% |
| 15 | glm-5.3-flash 6 providers | Friendli AI | 50.4B | $0.0060 | 99.64% |
| 16 | north-mini-code:free 1 provider | Cohere | 47.3B | $0 | 77.09% |
| 17 | gpt-6.1-terra 4 providers | OpenAI | 32.9B | $0.0232 | 99.99% |
| 18 | nemotron-3-nano-omni-30b-a3b-reasoning:free 1 provider | NVIDIA | 18.2B | $0 | 65.21% |
| 19 | claude-sonnet-5.5 3 providers | 18.0B | $0.0629 | 99.44% | |
| 20 | muse-spark-1.3-contributor 1 provider | Meta | 17.7B | $0.0011 | 99.00% |
Primary provider. The provider with the highest token volume for a model. Success rate equals 100% minus observed error rate. Costs are average observed request costs during the reporting period; zero indicates free traffic in the source data.
03 / Provider routing detail
Cost and reliability
Provider-level traffic share, average request cost, and success rate for the ten most-used models. Traffic share uses the percentage published by Kilo Gateway.
01
1.7T tokensnemotron-3.5-super-vl:free
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| NVIDIA | 100.00% | $0 | 95.45% |
02
1.2T tokensnemotron-3-ultra-550b-a55b:free
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| NVIDIA | 100.00% | $0 | 92.90% |
03
1.1T tokensbeam-free
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| Reflection | 100.00% | $0 | 93.99% |
04
567.6B tokensglyph-cluster
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| Stealth | 100.00% | $0 | 96.24% |
05
487.9B tokensdots-3-note-preview:free
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| Atlas Cloud | 100.00% | $0 | 87.96% |
03 / Provider routing detail
Cost and reliability
Provider-level traffic share, average request cost, and success rate for the ten most-used models. Traffic share uses the percentage published by Kilo Gateway.
06
258.9B tokensdeepseek-v4.1-flash-1017
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| Novita AI | 76.56% | $0.0038 | 99.82% |
| Alibaba | 13.28% | $0.0064 | 99.85% |
| Baseten | 5.84% | $0.0042 | 99.98% |
| Fireworks AI | 2.27% | $0.0047 | 99.94% |
| Together AI | 1.65% | $0.0035 | 99.94% |
| Relace | 0.17% | $0.0026 | 100.00% |
| DeepInfra | 0.13% | $0.00051 | 99.95% |
| Subconscious | 0.10% | $0.0045 | 100.00% |
07
151.4B tokenslaguna-s-2.1:free
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| Poolside | 100.00% | $0 | 96.57% |
03 / Provider routing detail
Cost and reliability
Provider-level traffic share, average request cost, and success rate for the ten most-used models. Traffic share uses the percentage published by Kilo Gateway.
08
125.1B tokensdeepseek-flash-latest
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| Novita AI | 84.59% | $0.0025 | 88.02% |
| Alibaba | 15.19% | $0.0053 | 99.14% |
| Baseten | 0.14% | $0.0061 | 99.84% |
| Fireworks AI | 0.07% | $0.0071 | 99.10% |
| Subconscious | 0.02% | $0.0073 | 100.00% |
09
123.6B tokensnemotron-3.5-lightning:free
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| NVIDIA | 100.00% | $0 | 82.04% |
10
104.6B tokensdeepseek-v4.1-flash
| Provider | Traffic share | Cost / request | Success rate |
|---|---|---|---|
| Novita AI | 92.62% | $0.0025 | 99.49% |
| Alibaba | 6.19% | $0.0088 | 99.92% |
| 1.18% | $0.0294 | 99.69% |