The Inference Report
May 15, 2026
The Value FrontierDaily rankings from SWE-rebench, a benchmark designed to fairly compare LLM capabilities on real-world software engineering tasks. Unlike other evaluations, it uses a standardized scaffolding for all models, continuously updates its dataset to prevent contamination, and runs each model five times to account for stochastic variance.
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 65.3% |
| 2 | gpt-5.2-2025-12-11-medium | 64.4% |
| 3 | GLM-5 | 62.8% |
| 4 | Junie | 62.8% |
| 5 | gpt-5.4-2026-03-05-medium | 62.8% |
| 6 | GLM-5.1 | 62.7% |
| 7 | Gemini 3.1 Pro Preview | 62.3% |
| 8 | DeepSeek-V3.2 | 60.9% |
| 9 | Claude Sonnet 4.6 | 60.7% |
| 10 | Claude Sonnet 4.5 | 60.0% |
Artificial Analysis composite index across coding, math, and reasoning benchmarks.
| # | Model | Score | tok/s | $/1M |
|---|---|---|---|---|
| 1 | GPT-5.5 | 60.2 | 66 | $11.25 |
| 2 | Claude Opus 4.7 | 57.3 | 62 | $10.94 |
| 3 | Gemini 3.1 Pro Preview | 57.2 | 126 | $4.50 |
| 4 | GPT-5.4 | 56.8 | 83 | $5.63 |
| 5 | Kimi K2.6 | 53.9 | 43 | $1.71 |
| 6 | MiMo-V2.5-Pro | 53.8 | 54 | $1.50 |
| 7 | GPT-5.3 Codex | 53.6 | 77 | $4.81 |
| 8 | Grok 4.3 | 53.2 | 85 | $1.56 |
| 9 | Claude Opus 4.6 | 52.9 | 50 | $10.94 |
| 10 | Muse Spark | 52.2 | 0 | $0.00 |
Output tokens per second — higher is faster. Minimum intelligence score of 40.
| # | Model | tok/s |
|---|---|---|
| 1 | Gemini 3 Flash Preview | 202 |
| 2 | GPT-5 Codex | 192 |
| 3 | Qwen3.6 35B A3B | 188 |
| 4 | GPT-5.1 Codex | 175 |
| 5 | GPT-5.4 mini | 164 |
| 6 | GPT-5.4 nano | 149 |
| 7 | Qwen3.5 122B A10B | 149 |
| 8 | MiMo-V2-Flash | 146 |
| 9 | Gemini 3 Pro Preview | 133 |
| 10 | Gemini 3.1 Pro Preview | 126 |
Blended cost per 1M tokens (3:1 input/output) — lower is cheaper. Minimum intelligence score of 40.
| # | Model | $/1M |
|---|---|---|
| 1 | MiMo-V2-Flash | $0.15 |
| 2 | DeepSeek V4 Flash | $0.175 |
| 3 | Hy3-preview | $0.20 |
| 4 | DeepSeek V3.2 | $0.337 |
| 5 | GPT-5.4 nano | $0.463 |
| 6 | MiniMax-M2.7 | $0.525 |
| 7 | KAT Coder Pro V2 | $0.525 |
| 8 | MiniMax-M2.5 | $0.525 |
| 9 | Qwen3.6 35B A3B | $0.557 |
| 10 | GPT-5 mini | $0.688 |