Benchmark explorer.
LiveBench Reasoning2.0 · 2026-10-14 (dataset 2026-09-02) · Higher is better
Source ↗Method
Single-run mean task score, 0–100.
Score × cost
Where performance meets price
LiveBench Reasoning · 1M input + 250K output tokens · uncached USD
21 priced resultsPareto frontier
Google: Gemini 4 Argon ↗95.1% LiveBench Reasoning · $2.25 workload · Google AI Studio · via OpenRouter
Evaluation: Effort: high · single run. Price: $1.00 input / $5.00 output per 1M tokens · google-ai-studio/flex · tier=flex.
Numbers match the model key. Shared positions show an index range.
Model key Score / workload
- NVIDIA: Nemotron 3.5 Lightning42.9% / $0.055
- OpenAI: GPT-6 Luna81.8% / $0.1125
- DeepSeek: DeepSeek V4.1 Flash86.7% / $0.135
- GLM-5.3-Flash77.6% / $0.1375
- Qwen3.8-27B80% / $0.145
- Google: Gemma 4 31B68.3% / $0.185
- Qwen: Qwen3.8 Flash87.4% / $0.2675
- DeepSeek-V4-Flash-Vision-Exp85.4% / $0.3773
- Reflection: Beam 1.0 501B A23B80.2% / $0.50
- Xiaomi: MiMo-V2.6-Pro79.6% / $0.6475
- GLM-5.385.8% / $0.8763
- Step-5-Preview83.8% / $1.675
- Google: Gemini 4 Argon95.1% / $2.25
- OpenAI: GPT-6.1 Sol92.6% / $2.25
- Meta: Muse Spark Open73.7% / $2.3125
- Kimi-K390.7% / $2.529
- Qwen: Qwen3.8 2.4T A95B88.2% / $3.50
- Anthropic: Claude Sonnet 5.591.6% / $4.50
- Anthropic: Claude Opus 5.592.2% / $9.00
- OpenAI: GPT-6 Astra92.7% / $11.25
- Anthropic: Claude Fable 5.191.7% / $22.50
Pareto frontier
- NVIDIA: Nemotron 3.5 LightningScore42.9%Workload cost$0.055Darkbloom · via OpenRouter · darkbloom/int4 · quantization=int4Evaluation: Effort: max · single run. Price: $0.025 input / $0.12 output per 1M tokens · darkbloom/int4 · quantization=int4.
- OpenAI: GPT-6 LunaScore81.8%Workload cost$0.1125OpenAI · via OpenRouter · openai/flex · tier=flex · prompt ≤ 272000 tokensEvaluation: Effort: max · single run. Price: $0.05 input / $0.25 output per 1M tokens · openai/flex · tier=flex · prompt ≤ 272000 tokens.
- DeepSeek: DeepSeek V4.1 FlashScore86.7%Workload cost$0.135Decart · via OpenRouter · decart/fp4 · quantization=fp4Evaluation: Effort: max · single run. Price: $0.09 input / $0.18 output per 1M tokens · decart/fp4 · quantization=fp4.
- Qwen: Qwen3.8 FlashScore87.4%Workload cost$0.2675Alibaba · via OpenRouter · alibaba/us-east-1Evaluation: Effort: max · single run. Price: $0.15 input / $0.47 output per 1M tokens · alibaba/us-east-1.
- Google: Gemini 4 ArgonScore95.1%Workload cost$2.25Google AI Studio · via OpenRouter · google-ai-studio/flex · tier=flexEvaluation: Effort: high · single run. Price: $1.00 input / $5.00 output per 1M tokens · google-ai-studio/flex · tier=flex.
Frontier: no result is cheaper with an equal or better score, or equally priced with a better score. Each point uses its lowest generally available workload price from one endpoint; evaluation and provider configurations remain visible.
| # | Model | LiveBench Reasoning | Configuration | API rate / unit | Source |
|---|---|---|---|---|---|
| 1 | Google: Gemini 4 Argongoogle | 95.1% | Effort: high · single runIndependent result | $1.00 in / $5.00 out/ 1M tokens · Google AI Studio · via OpenRoutergoogle-ai-studio/flex · tier=flex | Result ↗Oct 17, 2026 |
| 2 | OpenAI: GPT-6 Astraopenai | 92.7% | Effort: max · single runIndependent result | $5.00 in / $25.00 out/ 1M tokens · OpenAI · via OpenRouteropenai/flex · tier=flex · prompt ≤ 272000 tokens | Result ↗Oct 17, 2026 |
| 3 | OpenAI: GPT-6.1 Solopenai | 92.6% | Effort: max · single runIndependent result | $1.00 in / $5.00 out/ 1M tokens · OpenAI · via OpenRouteropenai/flex · tier=flex · prompt ≤ 272000 tokens | Result ↗Oct 17, 2026 |
| 4 | Anthropic: Claude Opus 5.5anthropic | 92.2% | Effort: max · single runIndependent result | $4.00 in / $20.00 out/ 1M tokens · Amazon Bedrock · via OpenRouteramazon-bedrock | Result ↗Oct 17, 2026 |
| 5 | Anthropic: Claude Fable 5.1anthropic | 91.7% | Effort: max · single runIndependent result | $10.00 in / $50.00 out/ 1M tokens · Amazon Bedrock · via OpenRouteramazon-bedrock | Result ↗Oct 17, 2026 |
| 6 | GPT-5.6 SolOpenAI | 91.7% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 7 | Anthropic: Claude Sonnet 5.5anthropic | 91.6% | Effort: max · single runIndependent result | $2.00 in / $10.00 out/ 1M tokens · Amazon Bedrock · via OpenRouteramazon-bedrock | Result ↗Oct 17, 2026 |
| 8 | Claude Opus 5Anthropic | 91.2% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 9 | Kimi-K3moonshotai | 90.7% | Effort: max · single runIndependent result | $0.27 in / $14.00 out/ 1M tokens · Relace · via OpenRouterrelace/fp4 · quantization=fp4 | Result ↗Oct 17, 2026 |
| 10 | GPT-5.6 TerraOpenAI | 90.6% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 11 | Muse Wonder 1.0Meta | 90.6% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 12 | Grok 4.6xAI | 90.5% | Effort: xhigh · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 13 | Muse Spark 1.2Meta | 90% | Effort: xhigh · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 14 | Claude Fable 5Anthropic | 89.7% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 15 | GPT-5.5OpenAI | 89.7% | Effort: xhigh · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 16 | Muse Spark 1.3Meta | 89.7% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 17 | Gemini 3.8 FlashGoogle | 89.3% | Effort: high · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 18 | Claude Opus 4.8Anthropic | 89.2% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 19 | Claude Sonnet 5Anthropic | 88.7% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 20 | GPT-6 SolOpenAI | 88.7% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 21 | Qwen: Qwen3.8 2.4T A95Bqwen | 88.2% | Effort: max · single runIndependent result | $2.00 in / $6.00 out/ 1M tokens · Alibaba · via OpenRouteralibaba | Result ↗Oct 17, 2026 |
| 22 | Gemini 3.7 FlashGoogle | 87.8% | Effort: high · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 23 | Qwen: Qwen3.8 Flashqwen | 87.4% | Effort: max · single runIndependent result | $0.15 in / $0.47 out/ 1M tokens · Alibaba · via OpenRouteralibaba/us-east-1 | Result ↗Oct 17, 2026 |
| 24 | Grok 4.5xAI | 87.2% | Effort: xhigh · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 25 | DeepSeek: DeepSeek V4.1 Flashdeepseek | 86.7% | Effort: max · single runIndependent result | $0.09 in / $0.18 out/ 1M tokens · Decart · via OpenRouterdecart/fp4 · quantization=fp4 | Result ↗Oct 17, 2026 |
| 26 | Claude Sonnet 4.6Anthropic | 86.4% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 27 | smaug-flashUnknown | 86.2% | Effort not specified · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 28 | GLM-5.3zai-org | 85.8% | Effort: max · single runIndependent result | $0.0632 in / $4.40 out/ 1M tokens · Relace · via OpenRouterrelace | Result ↗Oct 17, 2026 |
| 29 | GPT-5.6 LunaOpenAI | 85.6% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 30 | DeepSeek-V4-Flash-Vision-Expdeepseek-ai | 85.4% | Effort: max · single runIndependent result | $0.2156 in / $0.6468 out/ 1M tokens · DeepInfra · via OpenRouterdeepinfra/fp8 · quantization=fp8 | Result ↗Oct 17, 2026 |
| 31 | Claude Haiku 5.5Anthropic | 84.6% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 32 | Gemini 3.1 ProGoogle | 84% | Effort: high · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 33 | Mistral Large 4Mistral AI | 83.9% | Effort: high · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 34 | Step-5-Previewstepfun-ai | 83.8% | Effort: max · single runIndependent result | $1.00 in / $2.70 out/ 1M tokens · DeepInfra | Result ↗Oct 17, 2026 |
| 35 | Grok 4.7xAI | 82.7% | Effort: xhigh · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 36 | OpenAI: GPT-6 Lunaopenai | 81.8% | Effort: max · single runIndependent result | $0.05 in / $0.25 out/ 1M tokens · OpenAI · via OpenRouteropenai/flex · tier=flex · prompt ≤ 272000 tokens | Result ↗Oct 17, 2026 |
| 37 | Reflection: Beam 1.0 501B A23Breflection | 80.2% | Effort: max · single runIndependent result | $0.25 in / $1.00 out/ 1M tokens · Nebius · via OpenRouternebius/fp8 · quantization=fp8 | Result ↗Oct 17, 2026 |
| 38 | Qwen3.8-27BQwen | 80% | Effort: max · single runIndependent result | $0.02 in / $0.50 out/ 1M tokens · Tenstorrent · via OpenRoutertenstorrent/fp4 · quantization=fp4 | Result ↗Oct 17, 2026 |
| 39 | Xiaomi: MiMo-V2.6-Proxiaomi | 79.6% | Effort: high · single runIndependent result | $0.43 in / $0.87 out/ 1M tokens · DeepInfra · via OpenRouterdeepinfra/fp8 · quantization=fp8 | Result ↗Oct 17, 2026 |
| 40 | Nemotron 3.5 Super VLNVIDIA | 78.8% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 41 | InklingThinking Machines | 78.3% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 42 | GLM-5.3-Flashzai-org | 77.6% | Effort: max · single runIndependent result | $0.06 in / $1.60 out/ 1M tokens · Reka · via OpenRouterreka | Result ↗Oct 17, 2026 |
| 43 | Nemotron 3 UltraNVIDIA | 74.7% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 44 | MiniMax M3MiniMax | 74.5% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 45 | Meta: Muse Spark Openmeta | 73.7% | Effort: max · single runIndependent result | $1.25 in / $4.25 out/ 1M tokens · baseten · via Hugging Face | Result ↗Oct 17, 2026 |
| 46 | Google: Gemma 4 31Bgoogle | 68.3% | Effort: max · single runIndependent result | $0.10 in / $0.34 out/ 1M tokens · CoreWeave · via OpenRoutercoreweave/fp4 · quantization=fp4 | Result ↗Oct 17, 2026 |
| 47 | Claude Haiku 4.5Anthropic | 61.7% | Effort: max · single runIndependent result | See provider quotes | Result ↗Oct 17, 2026 |
| 48 | NVIDIA: Nemotron 3.5 Lightningnvidia | 42.9% | Effort: max · single runIndependent result | $0.025 in / $0.12 out/ 1M tokens · Darkbloom · via OpenRouterdarkbloom/int4 · quantization=int4 | Result ↗Oct 17, 2026 |