The Inference Report
March 14, 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 Code | 52.9% |
| 2 | Junie | 52.1% |
| 3 | Claude Opus 4.6 | 51.7% |
| 4 | gpt-5.2-2025-12-11-xhigh | 51.7% |
| 5 | gpt-5.2-2025-12-11-medium | 51.0% |
| 6 | gpt-5.1-codex-max | 48.5% |
| 7 | Claude Sonnet 4.5 | 47.1% |
| 8 | Gemini 3 Pro Preview | 46.7% |
| 9 | Gemini 3 Flash Preview | 46.7% |
| 10 | gpt-5.2-codex | 45.0% |
Artificial Analysis composite index across coding, math, and reasoning benchmarks.
| # | Model | Score | tok/s | $/1M |
|---|---|---|---|---|
| 1 | Gemini 3.1 Pro Preview | 57.2 | 117 | $4.50 |
| 2 | GPT-5.4 | 57 | 84 | $5.63 |
| 3 | GPT-5.3 Codex | 54 | 64 | $4.81 |
| 4 | Claude Opus 4.6 | 53 | 55 | $10.00 |
| 5 | Claude Sonnet 4.6 | 51.7 | 60 | $6.00 |
| 6 | GPT-5.2 | 51.3 | 61 | $4.81 |
| 7 | GLM-5 | 49.8 | 62 | $1.55 |
| 8 | Claude Opus 4.5 | 49.7 | 58 | $10.00 |
| 9 | GPT-5.2 Codex | 49 | 78 | $4.81 |
| 10 | Grok 4.20 Beta 0309 | 48.5 | 263 | $3.00 |
Output tokens per second — higher is faster. Minimum intelligence score of 40.
| # | Model | tok/s |
|---|---|---|
| 1 | Grok 4.20 Beta 0309 | 263 |
| 2 | GPT-5 Codex | 171 |
| 3 | Gemini 3 Flash Preview | 164 |
| 4 | Qwen3.5 122B A10B | 145 |
| 5 | MiMo-V2-Flash | 129 |
| 6 | Gemini 3 Pro Preview | 119 |
| 7 | Gemini 3.1 Pro Preview | 117 |
| 8 | GPT-5.1 Codex | 99 |
| 9 | Qwen3.5 27B | 89 |
| 10 | GPT-5.4 | 84 |
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 V3.2 | $0.315 |
| 3 | MiniMax-M2.5 | $0.525 |
| 4 | GPT-5 mini | $0.688 |
| 5 | Qwen3.5 27B | $0.825 |
| 6 | GLM-4.7 | $1.00 |
| 7 | Kimi K2 Thinking | $1.07 |
| 8 | Qwen3.5 122B A10B | $1.10 |
| 9 | Gemini 3 Flash Preview | $1.13 |
| 10 | Kimi K2.5 | $1.20 |