LLMs TigerLive LLM benchmarks

Mistral Medium 3.5: benchmark scores, price and ranking

Mistral Medium 3.5 from Mistral AI ranks #674 of 743 LLMs on the LLMs Tiger aggregate benchmark score (43.5), based on 11 benchmark results captured up to 2026-10-13.

Organisation
Mistral AI
Type
Frontier model
Open weights
No
Released
2026-04-28 (source: epoch)
Cheapest API price
$1.50 in / $7.50 out per million tokens (via azure_ai, observed 2026-10-13T19:47Z)
Aggregate rank
#674 of 743 (score 43.5)

Relative to other models, Mistral Medium 3.5 is strongest on Epoch ECI (#46 of 204) and weakest on FrontierCode (#15 of 15).

Across the 11 benchmarks where at least ten models are scored, Mistral Medium 3.5 is in the top 10% on 0 and in the top half on 3.

Its cheapest listed input price, $1.50 per million tokens, is lower than that of 25% of the 427 models with a listed price.

Mistral Medium 3.5 benchmark results

Mistral Medium 3.5 benchmark scores, ranked against all models measured on each benchmark
BenchmarkScoreRankSourceCaptured
Epoch ECI141#46 of 204Epoch AI Benchmarking Hub2026-10-13
WeirdML43.7%#45 of 109Epoch AI Benchmarking Hub2026-10-13
DTBench75.5%#66 of 146Epoch AI Benchmarking Hub2026-10-13
ALE-Bench764#40 of 79Epoch AI Benchmarking Hub2026-10-13
LMCA26.1%#62 of 108Epoch AI Benchmarking Hub2026-10-13
SciCode40.2%#64 of 106Epoch AI Benchmarking Hub2026-10-13
ProofBench9.0%#38 of 46Epoch AI Benchmarking Hub2026-10-13
Surface Evolver26.9%#11 of 13Epoch AI Benchmarking Hub2026-10-13
WebDev Arena1,265#57 of 67Epoch AI Benchmarking Hub2026-10-13
CritPt0.0%#106 of 115Epoch AI Benchmarking Hub2026-10-13
FrontierCode8.0%#15 of 15Epoch AI Benchmarking Hub2026-10-13

About Mistral Medium 3.5

Mistral Medium 3.5 is a frontier model from Mistral AI. It was released on 2026-04-28, according to Epoch AI. The cheapest API listing we track charges $1.50 per million input tokens and $7.50 per million output tokens, through azure_ai. We hold 11 benchmark results for it across 4 areas, and none is in the top tenth of its board.

Where Mistral Medium 3.5 stands

On our aggregate score, Mistral Medium 3.5 is #674 of 743 ranked models, in the lower half. Its score is 43.5; the scale is centred on 50, so it sits 6.5 points below the typical model. That beats 9% of the ranked field. Among the 44 ranked Mistral AI models it is #35. The next one up is Mixtral-8x22B-Instruct-v0.1 at 43.5. The next one down is mistral-large-2407 at 43.3.

Mistral Medium 3.5, benchmark by benchmark

Epoch ECI puts Mistral Medium 3.5 at 141, #46 of 204. That beats 77% of the models tested, but 15 points short of Qwen 3.8 Max (156). The results either side are Gemini 2.5 Flash (May 2025) (142) and DeepSeek-R1 (May 2025) (141). It covers Epoch Capabilities Index: Epoch's own aggregate of ~40 benchmarks. The board median is 127, so Mistral Medium 3.5 is 14 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

On WeirdML, Mistral Medium 3.5 reaches 43.7%. It ranks #45 of 109, ahead of 59% of the field, and 49.6 points behind GPT-6 Astra (max) at 93.3%. The results either side are Claude Opus 4 (43.7%) and o3-mini (high) (43.7%). It covers unusual machine-learning tasks. The board median is 41.2%, so Mistral Medium 3.5 is 2.5 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Mistral Medium 3.5 manages 75.5% on DTBench, good for #66 of 146. Claude Opus 5.5 (max) leads at 98.9%, 23.5 points further up, and 55% of models sit below. The results either side are qwen3-235b-a22b (75.7%) and Gemma 4 26B A4B (74.9%). It covers dynamic tool-use benchmark. The board median is 68.8%, so Mistral Medium 3.5 is 6.7 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Mistral Medium 3.5 lands below the middle on ALE-Bench: 764, #40 of 79, with 49% of the field scoring higher. GPT-6 Astra (max) leads at 2,951; the closest result above is GLM-5 at 766. It covers AtCoder heuristic-optimisation contests (performance rating). The median score on this board is also 764. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Mistral Medium 3.5 lands below the middle on LMCA: 26.1%, #62 of 108, with 56% of the field scoring higher. Claude Opus 5.5 (max) leads at 68.2%; the closest result above is o4-mini (high) at 26.5%. It covers language-model capability assessment. The board median is 29.2%, so Mistral Medium 3.5 is 3.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Mistral Medium 3.5 lands below the middle on SciCode: 40.2%, #64 of 106, with 59% of the field scoring higher. Claude Opus 5.5 (max) leads at 66.9%; the closest result above is nemotron-3-ultra at 40.3%. It covers scientific research code. The board median is 44.3%, so Mistral Medium 3.5 is 4.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Surface Evolver is a weak spot. Mistral Medium 3.5 scores 26.9%, #11 of 13, behind 77% of the models measured. The leader, Kimi K3 (max), reaches 93.0%. It covers geometry optimisation with a solver. The board median is 46.2%, so Mistral Medium 3.5 is 19.4 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

WebDev Arena is a weak spot. Mistral Medium 3.5 scores 1,265, #57 of 67, behind 84% of the models measured. The leader, Claude Opus 5.5 (max), reaches 1,820. It covers building web apps, crowd-voted (Elo). The board median is 1,433, so Mistral Medium 3.5 is 169 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

CritPt is a weak spot. Mistral Medium 3.5 scores 0.0%, #106 of 115, behind 91% of the models measured. The leader, GPT-5.6 Sol (max), reaches 32.3%. It covers unpublished research-level physics problems. The board median is 2.9%, so Mistral Medium 3.5 is 2.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

FrontierCode is a weak spot. Mistral Medium 3.5 scores 8.0%, #15 of 15, behind 93% of the models measured. The leader, Claude Opus 5 (max), reaches 53.4%. It covers hard programming tasks. The board median is 31.8%, so Mistral Medium 3.5 is 23.8 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

The full table above also lists the other 1 result, including boards with fewer than ten models, where a rank says little.

Strengths and weaknesses

Software engineering: weak, ahead of 26% of the field on average across 4 benchmarks. Its best board here is ALE-Bench (#40 of 79) and its weakest is FrontierCode (#15 of 15). Reasoning, maths and science: weak, ahead of 13% of the field on average across 2 benchmarks. Its best board here is ProofBench (#38 of 46) and its weakest is CritPt (#106 of 115). General knowledge and chat: uneven, ahead of 49% of the field on average across 4 benchmarks. Its best board here is Epoch ECI (#46 of 204) and its weakest is Surface Evolver (#11 of 13).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Mistral Medium 3.5 costs more than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on GPT-4.1 ($2.00) and mistral-large-2407 ($2.00).

Release history

The Mistral AI model released before it was Mistral Small 4 on 2026-03-16, which scores 46.5. The next one was Mistral Large 4 (high) on 2026-10-06, at 47.7.

Reading these numbers

The 11 scores on this page come from 1 source: Epoch AI Benchmarking Hub (11). No official leaderboard supplied any of them, so treat them as third-party measurements.

Models ranked nearby

More from Mistral AI

Compare it on the live leaderboard → · How the aggregate score works