Compare/Mistral Large (Feb '24) vs GPT-5.3 Codex (xhigh)

Mistral Large (Feb '24)vsGPT-5.3 Codex (xhigh)

Side-by-side comparison of pricing, 12 benchmarks, and generation speed.

Mistral

Mistral Large (Feb '24)

Input
$4/M
Output
$12/M
Speed
TTFT
OpenAI

GPT-5.3 Codex (xhigh)

Input
$1.75/M
Output
$14/M
Speed
81 tok/s
TTFT
68.19s

Winner by Category

Cheaper
GPT-5.3 Codex (xhigh)
Faster (tok/s)
GPT-5.3 Codex (xhigh)
Lower Latency
Mistral Large (Feb '24)
Benchmarks (3-7)
GPT-5.3 Codex (xhigh)

Pricing Comparison

MetricMistral Large (Feb '24)GPT-5.3 Codex (xhigh)
Input ($/M tokens)$4$1.75
Output ($/M tokens)$12$14
Cost for 1M input + 100K output tokens:
Mistral Large (Feb '24)$5.20
GPT-5.3 Codex (xhigh)$3.15

Speed Comparison

Output Speed (tokens/s) — higher is better
Mistral Large (Feb '24)
GPT-5.3 Codex (xhigh)
81 tok/s
Time to First Token (seconds) — lower is better
Mistral Large (Feb '24)
GPT-5.3 Codex (xhigh)
68.19s

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
9.954.0
Coding Index
53.1
Math Index
GPQA Diamond
35.1%91.5%
MMLU-Pro
51.5%
LiveCodeBench
17.8%
AIME 2025
MATH-500
52.7%
Humanity's Last Exam
3.4%39.9%
SciCode
20.8%53.2%
IFBench
75.4%
TerminalBench
53.0%
Mistral Large (Feb '24)3 wins
7 winsGPT-5.3 Codex (xhigh)

Frequently Asked Questions

Which is cheaper, Mistral Large (Feb '24) or GPT-5.3 Codex (xhigh)?

GPT-5.3 Codex (xhigh) is cheaper overall. Its blended price (3:1 input/output ratio) is $4.81/M tokens vs $6.00/M for Mistral Large (Feb '24).

Which model performs better on benchmarks?

GPT-5.3 Codex (xhigh) wins 7 out of 12 benchmarks compared to 3 for Mistral Large (Feb '24). See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

GPT-5.3 Codex (xhigh) generates tokens faster at 81 tok/s vs 0 tok/s. Mistral Large (Feb '24) also has lower time-to-first-token (0.00s vs 68.19s).

When should I use Mistral Large (Feb '24) vs GPT-5.3 Codex (xhigh)?

Choose based on your priorities: GPT-5.3 Codex (xhigh) for lower cost, GPT-5.3 Codex (xhigh) for stronger benchmark performance, and GPT-5.3 Codex (xhigh) for faster generation. For latency-sensitive apps, check the TTFT comparison above.